diff --git a/.gitignore b/.gitignore
index f6ccf6b..5b8858a 100644
--- a/.gitignore
+++ b/.gitignore
@@ -236,4 +236,16 @@ data
 todo.md
 
 
+# ignoring trash file
+trash.py
+
+
+# ignoring jupyter notebook
+explore.ipynb
+
+
+# ignoring jupyter notebook
+tests/playground.py
+
+tests/.pytest_cache
 .pytest_cache
\ No newline at end of file
diff --git a/ConversationAnalyzer.py b/ConversationAnalyzer.py
new file mode 100644
index 0000000..fe30520
--- /dev/null
+++ b/ConversationAnalyzer.py
@@ -0,0 +1,152 @@
+import pandas as pd
+from utils import date_checker, period_checker, subject_checker, generate_time_series, get_stats_for_intervals
+
+
+class ConversationAnalyzer:
+    def __new__(cls, name, messages, *args, **kwargs):
+        if messages is None:  # This deals with the case if no messages
+            return None
+        return super(ConversationAnalyzer, cls).__new__(cls, *args, **kwargs)
+
+    def __init__(self, name, messages):
+        self.name = name
+        self.df = messages
+
+    def __str__(self):
+        return f'{self.name}: {list(self.df.index)}'
+
+    @property
+    def stats(self):
+        return self.get_stats(self.df)
+
+    # TODO has to be tested
+    def get_time_series_data(self, subject='all', **kwargs):
+        time_series = generate_time_series(**kwargs)
+        return get_stats_for_intervals(self.get_stats, self.df, time_series, subject=subject)
+
+    def get_plotable_time_series_data(self, interval_stats, statistic):
+        for k, v in interval_stats.items():
+            if isinstance(v, ConversationStats):
+                interval_stats[k] = getattr(v, statistic)
+        return interval_stats
+
+    def get_stats(self, df, subject='all', start=None, end=None, period=None):
+        df = self.filter_by_input(df, subject=subject, start=start, end=end, period=period)
+        stats = ConversationStats(df)
+        return stats
+
+    @staticmethod
+    @subject_checker
+    @date_checker
+    @period_checker
+    def filter_by_input(df, subject='all', start=None, end=None, period=None):
+        if subject == 'me':
+            df = df[df.sender_name == 'Levente Csőke']
+        elif subject == 'partner':
+            df = df[df.sender_name != 'Levente Csőke']
+        if start and end:
+            df = df.loc[start:end]
+        elif start and not end:
+            df = df.loc[start:start + period]
+        elif not start and end:
+            df = df.loc[end - period:end]
+        return df
+
+
+class ConversationStats:
+    """
+    Statistics of conversation with one person.
+    """
+
+    # TODO do we need this or not?!?! smh
+    # def __new__(cls, df, *args, **kwargs):
+    #     if not len(df.index):  # This deals with the case if input df is empty
+    #         return None
+    #     return super(ConversationStats, cls).__new__(cls, *args, **kwargs)
+
+    def __init__(self, df):
+        self.df = df
+
+    def __repr__(self):
+        return f'{self.msg_count}'
+
+    @property
+    def messages(self):
+        return self.df.content.dropna()
+
+    @property
+    def words(self):
+        return self.get_words()
+
+    # 1.
+    @property
+    def msg_count(self):
+        return len(self.df)
+
+    # 2.
+    @property
+    def unique_msg_count(self):
+        return len(self.messages.unique())
+
+    # 3.
+    @property
+    def most_used_msgs(self):
+        # TODO first few (1-10) messages
+        return self.messages.value_counts()
+
+    # 4.
+    @property
+    def msg_frequency(self):
+        # TODO this has been most likely depracated
+        pass
+
+    # 5.
+    @property
+    def word_count(self):
+        return len(self.words)
+
+    # 6.
+    @property
+    def unique_word_count(self):
+        return len(set(self.words))
+
+    # 7.
+    @property
+    def most_used_words(self):
+        s = pd.Series(self.words)
+        return s.value_counts()
+
+    # 8.
+    @property
+    def word_frequency(self):
+        pass
+
+    # 9.
+    @property
+    def char_count(self):
+        char_count = 0
+        for word in self.words:
+            char_count += len(word)
+        return char_count
+
+    # 10.
+    @property
+    def most_used_chars(self):
+        return None  # TODO or not  https://stackoverflow.com/questions/4131123/finding-the-most-frequent-character-in-a-string
+
+    # 11.
+    @property
+    def rate_of_media_messages(self):
+        pass  # TODO what?
+
+    def get_words(self):
+        token_list = self.messages.str.lower().str.split()
+        words = []
+        for tokens in token_list:
+            # print(tokens)
+            if not isinstance(tokens, list):
+                print('WARNING! Not a list!')
+                continue  # TODO ??? check this
+            for token in tokens:
+                words.append(token)
+        return words
diff --git a/Conversations.py b/Conversations.py
new file mode 100644
index 0000000..112f336
--- /dev/null
+++ b/Conversations.py
@@ -0,0 +1,116 @@
+import os
+from FacebookData import FacebookData
+import pandas as pd
+
+from datetime import datetime
+
+MESSAGE_SUBPATH = 'messages/inbox'
+
+
+class Conversations:
+    def __init__(self, data_path):
+        self.data_path = f'{data_path}/{MESSAGE_SUBPATH}'
+
+    def get_people(self):
+        json_paths = self.walk_directory_and_search('.json')
+        return self.extract_names_from_convos(json_paths)
+
+    def walk_directory_and_search(self, extension):
+        paths = []
+        for root, dirs, files in os.walk(self.data_path):
+            for name in files:
+                if name.endswith(extension):
+                    paths.append(os.path.join(root, name))
+        return paths
+
+    # TODO simplify this function!! also this takes very long
+    @staticmethod
+    def extract_names_from_convos(jsons):
+        name_data_map = {}
+        count = 0
+        for file in jsons:
+            msg = Messages(file)
+            for participant in msg.participants:
+                key = participant if msg.ttype == 'Regular' else f'group_{count}'
+                if key == 'Facebook User':  # TODO ?? what to do with this??
+                    continue
+                if name_data_map.get(key) and key.startswith(
+                        'group'):  # making sure run only once even if it is a group
+                    continue
+                if name_data_map.get(key):
+                    dfs = [name_data_map[key]['messages'], msg.df]
+                    name_data_map[key]['messages'] = pd.concat(dfs, ignore_index=False).sort_index()
+                else:
+                    name_data_map[key] = {
+                        'title': msg.title,
+                        'compact_name': msg.compact_names,  # TODO is list ok for if length is  only  1??
+                        # 'participants': msg.participants + ['Levente Csőke'],
+                        'participants': msg.participants,
+                        'messages': msg.df,
+                        'friend': None,
+                        'messages_dir': msg.messages_dir,
+                        'media_dir': msg.media_dir
+                    }
+            if msg.ttype == 'RegularGroup':
+                count += 1
+
+        return name_data_map
+
+
+class Messages(FacebookData):
+    def __init__(self, json_path):
+        super().__init__(json_path)
+        self.to_df()
+        self.set_date_as_index()
+
+    def to_df(self):
+        self._df = pd.DataFrame(self.decoded.get('messages'))
+
+    def set_date_as_index(self):
+        # TODO maybe not needed; could calculate real time
+        date_series = self._df.timestamp_ms.apply(self.ts_to_date)
+        self._df = self._df.set_index(date_series).iloc[::-1]
+
+    @property
+    def names(self):
+        return pd.DataFrame(self.participants)[0]
+
+    @property
+    def participants(self):
+        participants = self.decoded.get('participants')
+        # TODO I should be IN
+        # but this breaks stuff at TestMessagingAnalyzer
+        return [p.get('name') for p in participants if p.get('name') != 'Levente Csőke']
+        # return [p.get('name') for p in participants if p.get('name')]
+
+    @property
+    def title(self):
+        return self.decoded.get('title')
+
+    @property
+    def ttype(self):
+        return self.decoded.get('thread_type')
+
+    @property
+    def messages_dir(self):
+        thread_path = self.decoded.get('thread_path')
+        if not thread_path.startswith('inbox/'):
+            raise ValueError('Something is not okay.')
+        # TODO here or in the upper function where we extract names
+        return thread_path.split('/')[1].lower()
+
+    @property
+    def media_dir(self):
+        # todo what should the path contain
+        for media in ['photos', 'gifs', 'files', 'videos', 'audio']:
+            if media in self._df.columns:
+                media_in_msg = list(self._df[media][self._df[media].notnull()])
+                # if len(media_in_msg) > 1:  # TODO is this ok. i think it is. think multiple photos sent once
+                #    print('Media in msg is bigger than 1')
+                uri = media_in_msg[0][0].get('uri')
+                return os.path.dirname(os.path.dirname(uri))
+        return None
+
+    @staticmethod
+    def ts_to_date(date):
+        return datetime.fromtimestamp(date / 1000)  # .strftime('%Y-%m-%d')
diff --git a/FacebookData.py b/FacebookData.py
new file mode 100644
index 0000000..a82c896
--- /dev/null
+++ b/FacebookData.py
@@ -0,0 +1,40 @@
+from utils import read_json, decode_text, accents_map
+
+
+class FacebookData:
+    def __init__(self, json_path):
+        self.json_path = json_path
+        self._df = None
+
+    @property
+    def df(self):
+        return self._df
+
+    @property
+    def decoded(self):
+        return decode_text(self.json)
+
+    @property
+    def json(self):
+        return read_json(self.json_path)
+
+    @property
+    def compact_names(self):
+        # NOTE this is the place where we change pd/np to builtin
+        # do we have to do this?
+        name_list = list(self.without_accent_and_whitespace(self.lower_names(self.names)))
+        return name_list[0] if len(name_list) == 1 else name_list
+
+    @staticmethod
+    def lower_names(col):
+        return col.str.lower()
+
+    @staticmethod
+    def without_accent_and_whitespace(col):
+        def replace_accents(text):
+            for char in accents_map.keys():
+                if char in text:
+                    text = text.replace(char, accents_map[char])
+            return text.replace(' ', '')
+
+        return col.apply(replace_accents)
diff --git a/Friends.py b/Friends.py
new file mode 100644
index 0000000..6e0e991
--- /dev/null
+++ b/Friends.py
@@ -0,0 +1,36 @@
+import pandas as pd
+import os
+from FacebookData import FacebookData
+from utils import accents_map
+
+
+class Friends(FacebookData):
+
+    def __init__(self, *args):
+        super().__init__(*args)
+
+        # self.path = 'data/friends'
+        # self.json_path = f'{self.path}/friends.json'
+
+        self.to_df()
+
+    def get_people(self):
+        names = {}
+        for name, compact in zip(self.names, self.compact_names):
+            names[name] = {
+                'title': name,
+                'compact_name': compact,
+                'messages': None,
+                'friend': True,
+                'participants': None,
+                'messages_dir': None,
+                'media_dir': None
+            }
+        return names
+
+    def to_df(self):
+        self._df = pd.DataFrame(self.decoded.get('friends'))
+
+    @property
+    def names(self):
+        return self.df.name
diff --git a/Group.py b/Group.py
new file mode 100644
index 0000000..41b966f
--- /dev/null
+++ b/Group.py
@@ -0,0 +1,47 @@
+# TODO groups should be searched by looking into jsons unfortunately :(
+# because of directory says others
+# maybe we dont use groups right away?
+
+
+class Group:
+    def __init__(self, name=None, title=None, messages=None, compact=None, messages_dir=None, media_dir=None,
+                 members=None):
+        self._name = name
+        self._title = title
+        self._messages = messages
+        self._compact_name = compact
+        self._messages_dir = messages_dir
+        self._media_dir = media_dir
+        self._members = members
+
+    @property
+    def name(self):
+        return self._name
+
+    @property
+    def title(self):
+        return self._title
+
+    @property
+    def messages(self):
+        return self._messages
+
+    # @property
+    # def get_message_jsons(self):
+    #     return self._messages
+
+    @property
+    def media_dir(self):
+        return self._media_dir
+
+    @property
+    def messages_dir(self):
+        return self._messages_dir
+
+    @property
+    def compact_name(self):
+        return self._compact_name
+
+    @property
+    def members(self):
+        return self._members
diff --git a/Individual.py b/Individual.py
new file mode 100644
index 0000000..a9f8d03
--- /dev/null
+++ b/Individual.py
@@ -0,0 +1,47 @@
+class Individual:
+    def __init__(self, name=None, title=None,compact=None, messages=None, friend=None, messages_dir=None, media_dir=None,
+                 member_of=None):
+        self._name = name
+        self._title = title
+        self._compact_name = compact
+        self._messages = messages
+        self._friend = friend
+        self._messages_dir = messages_dir
+        self._media_dir = media_dir
+        self._member_of = member_of
+
+
+    def __repr__(self):
+        return self.name
+
+    @property
+    def name(self):
+        return self._name
+
+    @property
+    def title(self):
+        return self._title
+
+    @property
+    def messages(self):
+        return self._messages
+
+    @property
+    def friend(self):
+        return self._friend
+
+    @property
+    def media_dir(self):
+        return self._media_dir
+
+    @property
+    def messages_dir(self):
+        return self._messages_dir
+
+    @property
+    def compact_name(self):
+        return self._compact_name
+
+    @property
+    def member_of(self):
+        return self._member_of
diff --git a/Me.py b/Me.py
new file mode 100644
index 0000000..3293bf7
--- /dev/null
+++ b/Me.py
@@ -0,0 +1,10 @@
+from FacebookData import FacebookData
+
+
+class Me(FacebookData):
+    def __init__(self, *args):
+        super().__init__(*args)
+
+    @property
+    def name(self):
+        return 'Levente Csőke'
diff --git a/MessagingAnalyzer.py b/MessagingAnalyzer.py
new file mode 100644
index 0000000..07965e5
--- /dev/null
+++ b/MessagingAnalyzer.py
@@ -0,0 +1,135 @@
+from utils import year_converter, month_converter, generate_time_series, get_stats_for_intervals
+from datetime import datetime, date, timedelta
+import pandas as pd
+from ConversationAnalyzer import ConversationAnalyzer
+
+"""
+
+"""
+
+
+class MessagingAnalyzer:
+    def __init__(self, names, people):
+        # TODO input people only. class ill know what to do
+        self.names = names
+        self.people = people
+
+    def time_series_analysis_for_all(self, subject=None, **kwargs):
+        time_series = generate_time_series(**kwargs)
+        stacked_df = self.stack_dfs(self.people)
+        interval_stats = get_stats_for_intervals(self.get_stats, stacked_df, time_series, subject=subject)
+
+    def get_stats(self, df, subject='all', start=None, end=None, period=None):
+        # TODO
+        # here you have to do something with it
+        pass
+
+    def get_count(self, attribute, subject='all', start=None, end=None, period=None):
+        count = 0
+        # we have a list of names we want to iterate over
+        for name in self.names:
+            stats = self.get_conversation_stats(name=name, subject=subject, start=start, end=end, period=period)
+            if stats is not None:  # TODO too explicit; needed because it is possible that None will be returned, if t got an empty df
+                count += getattr(stats, attribute)
+        return count
+
+    def get_conversation_stats(self, name, subject='all', start=None, end=None, period=None):
+        messages = self.people.get(name).messages
+        analyzer = ConversationAnalyzer(name, messages)
+        if analyzer is None:  # TODO this is too explicit ?!
+            return None
+        return analyzer.get_stats(messages, subject=subject, start=start, end=end, period=period)
+
+    def total_number_of_(self, attribute, subject='all', **kwargs):
+        return self.get_count(attribute=attribute, subject=subject, **kwargs)
+
+    # 1. Ranking of friends by total count of messages/words/characters (also by year/month/day/hour)
+    def total_number_of_messages(self, **kwargs):
+        return self.total_number_of_(attribute='msg_count', **kwargs)
+
+    def total_number_of_words(self, **kwargs):
+        return self.total_number_of_(attribute='word_count', **kwargs)
+
+    def total_number_of_characters(self, **kwargs):
+        return self.total_number_of_(attribute='char_count', **kwargs)
+
+    # 2. Ranking of friends who I sent the most messages/words/characters (also by year/month/day/hour)
+    def total_number_of_messages_sent(self, **kwargs):
+        return self.total_number_of_(attribute='msg_count', subject='me', **kwargs)
+
+    def total_number_of_words_sent(self, **kwargs):
+        return self.total_number_of_(attribute='word_count', subject='me', **kwargs)
+
+    def total_number_of_characters_sent(self, **kwargs):
+        return self.total_number_of_(attribute='char_count', subject='me', **kwargs)
+
+    # 3. Ranking of friends who sent the most messages/words/characters (also by year/month)
+    def total_number_of_messages_received(self, **kwargs):
+        return self.total_number_of_(attribute='msg_count', subject='partner', **kwargs)
+
+    def total_number_of_words_received(self, **kwargs):
+        return self.total_number_of_(attribute='word_count', subject='partner', **kwargs)
+
+    def total_number_of_characters_received(self, **kwargs):
+        return self.total_number_of_(attribute='char_count', subject='partner', **kwargs)
+
+    # 4. Most used messages/words in convos by me/partner (also by year/month/day/hour)
+    def most_used_messages_by_me(self, **kwargs):
+        """
+        >>> s1 = pd.Series([3, 1, 2, 3, 4, 1, 1])
+        >>> s2 = pd.Series([3, 2, 1, 1])
+        >>> s1_vc = s1.value_counts()
+        >>> s2_vc = s2.value_counts()
+        TODO (later) most used is already a problem:
+          - because its a series of all the unique messages/words ever used in a convo
+          - it contains strings like ':d', ':p' and 'xd'
+          - from all the convos the result of value_counts has to be cleared
+          and has to be truncated (that is not use the 200th most used word, only top10 let's say)
+          - then these series has to be merged in a way that the same string's counts are added up
+          - what about typos????!
+        """
+        pass
+
+    def most_used_messages_by_partners(self, **kwargs):
+        pass
+
+    def most_used_words_by_me(self, **kwargs):
+        pass
+
+    def most_used_words_by_partners(self, **kwargs):
+        pass
+
+    # 5. Number of messages sent/got on busiest period (by year/month/day/hour)
+    def days_when_most_messages_sent(self):
+        # TODO hard algorithmic problem
+        pass
+
+    def days_when_most_messages_received(self):
+        pass
+
+    def hours_when_most_messages_sent(self):
+        # TODO
+        # is this referring to the absolute hour most messages sent??
+        # like: 2014.07.25. 15h-16h
+        # OR
+        # the pattern of most messages sent between this and this hours
+        # like: 20h-21h
+        # ACTUALLY BOTH
+        # for years/months/days/hours
+        # BUT this comes from the time series analysis
+        pass
+
+    def hours_when_most_messages_received(self):
+        pass
+
+    # 6. Time series: dict of 'year/month/day/hour : number of messages/words/characters (also sent/got) for user/all convos'
+    # TODO
+
+    @staticmethod
+    def stack_dfs(people):
+        dfs = []
+        for data in people.values():
+            if data.messages is not None:
+                dfs.append(data.messages)
+        # TODO do I need to sort by index (date)? yes!
+        return pd.concat(dfs).sort_index()  # TODO why ignore_index??
diff --git a/Miner.py b/Miner.py
index e69de29..3b32806 100644
--- a/Miner.py
+++ b/Miner.py
@@ -0,0 +1,45 @@
+DATA_PATH = '/home/levente/projects/facebook-data-miner/data'
+
+from People import People
+from ConversationAnalyzer import ConversationAnalyzer
+from MessagingAnalyzer import MessagingAnalyzer
+
+
+class Miner:
+    def __init__(self):
+        pass
+
+    @staticmethod
+    def analyze_messages():
+        p = People(path=DATA_PATH)
+
+        stats = {}
+
+        for name, person in p.individuals.items():
+            #assert name == person.name, 'ERRRRRRROR!!!'
+            if person.messages is None:
+                stats[person.name] = None
+                continue
+            analyzer = ConversationAnalyzer(person.name, person.messages)
+            stats[person.name] = analyzer.stats
+            # if stats[person.name].get('message_count').get('me') > 5000:
+            #    top[person.name] = stats[person.name]
+        example = stats['Dániel Nagy']
+        print()
+
+        # print('LEN: ', len(top.keys()))
+        # top_all = {name: data.get('message_count').get('all') for name, data in top.items()}
+        # analyzer.visualize_stats(top)
+
+    @staticmethod
+    def analyze_messaging():
+        p = People(path=DATA_PATH)
+
+        msg_analyzer = MessagingAnalyzer(p.names, p.individuals)
+
+        msgs = msg_analyzer.total_number_of_messages()
+
+
+if __name__ == '__main__':
+    m = Miner()
+    m.analyze_messages()
diff --git a/People.py b/People.py
new file mode 100644
index 0000000..11d1887
--- /dev/null
+++ b/People.py
@@ -0,0 +1,79 @@
+from Individual import Individual
+from Conversations import Conversations
+from Friends import Friends
+
+# from Me import Me
+DATA_PATH = '/home/levente/projects/facebook-data-miner/data'
+import time
+from Group import Group
+
+
+# TODO we dont need both data and individuals... or??
+
+class People:
+    def __init__(self, path=None):
+        self.data_path = path if path else DATA_PATH
+        self._names = []
+        self._individuals = {}
+        self._groups = []
+        self._data = self.get_people()  # TODO is this supposed to be here or elsewhere
+        self.to_individuals()  # TODO is this supposed to be here or elsewhere
+
+    @property
+    def data(self):
+        return self._data
+
+    @property
+    def names(self):
+        return self._names
+
+    @property
+    def individuals(self):
+        return self._individuals
+
+    @property
+    def groups(self):
+        return self._groups
+
+    def get_people(self):
+        start = time.time()
+        friends = Friends(self.data_path + '/friends/friends.json')
+        people1 = friends.get_people()
+        print('friends: ', time.time() - start)
+
+        # TODO LATER too slow
+        # takes about 30 secs both
+        # read it once, store it in DB OR?
+        start = time.time()
+        conversations = Conversations(self.data_path)
+        people2 = conversations.get_people()
+        print('convos: ', time.time() - start)
+
+        return self.unify_people(people1, people2)
+
+    def to_individuals(self):  # TODO maybe rather split_convos or differentiate_convos
+        start = time.time()
+        for person, data in self._data.items():
+            if person.startswith('group'):
+                g = Group(name=data.get('name'), title=data.get('title'), messages=data.get('messages'),
+                          compact=data.get('compact_name'), messages_dir=data.get('messages_dir'),
+                          media_dir=data.get('media_dir'), members=None)
+                self._groups.append(g)
+            else:
+                indie = Individual(name=person, title=data.get('title'), messages=data.get('messages'),
+                                   compact=data.get('compact_name'), messages_dir=data.get('messages_dir'),
+                                   media_dir=data.get('media_dir'), member_of=None)
+                self._names.append(person)
+                self._individuals[person] = indie
+        print('indies: ', time.time() - start)
+
+    @staticmethod
+    def unify_people(friends, convos):
+        for person, data in friends.items():
+            if not convos.get(person):
+                convos[person] = data
+            convos[person]['friend'] = True
+        return convos
+
+# if __name__ == '__main__':
+#     p = People()
diff --git a/README.md b/README.md
index d9dbdd3..943a123 100644
--- a/README.md
+++ b/README.md
@@ -1 +1,14 @@
-# facebook-data-miner
\ No newline at end of file
+# Facebook Data Miner
+This repository has a purpose to provide a set of tools, with which one can analyze their facebook data locally.
+
+The codebase is under development. 
+
+Features will be added gradually, starting with basic analysis of the messages.
+Don't expect too fast development tho'.
+
+More info soon...
+
+## Contribution
+Help is more than welcome. If somebody feel the urge to contribute, I would share my plans with them.
+
+Ideas are welcome too. Feel free to open a new issue.
\ No newline at end of file
diff --git a/Visualizer.py b/Visualizer.py
new file mode 100644
index 0000000..b4a3779
--- /dev/null
+++ b/Visualizer.py
@@ -0,0 +1,36 @@
+import matplotlib.pyplot as plt
+import seaborn as sns
+import pandas as pd
+from People import People
+from ConversationAnalyzer import ConversationAnalyzer
+
+
+# plt.rcParams.update({'figure.figsize': (10, 7), 'figure.dpi': 120})
+
+class Visualizer:
+    def __init__(self):
+        pass
+
+    def plot_time_series(self, x, y, title="Time series", xlabel='Date', ylabel='Value', dpi=100):
+        plt.figure(figsize=(16, 5), dpi=dpi)
+        plt.plot(x, y, color='tab:red')
+        plt.gca().set(title=title, xlabel=xlabel, ylabel=ylabel)
+        plt.show()
+
+
+def set_up(people, name, interval='y'):
+    analyzer = ConversationAnalyzer(name, people.individuals.get(name).messages)
+    interval_stats = analyzer.get_time_series_data()
+    stats = interval_stats.get(interval)
+    return analyzer.get_plotable_time_series_data(stats, statistic='msg_count')
+
+
+if __name__ == "__main__":
+    v = Visualizer()
+    TEST_DATA_PATH = '/home/levente/projects/facebook-data-miner/tests/test_data'
+    people = People(path=TEST_DATA_PATH)
+    names = ['Teflon Musk', 'Tőke Hal']
+    for name in names:
+        data = set_up(people, name, interval='d')
+        df = pd.DataFrame(data.items(), columns=['date', 'value'])
+        v.plot_time_series(x=df.date, y=df.value, title=name)
diff --git a/tests/TestConversationAnalyzer.py b/tests/TestConversationAnalyzer.py
new file mode 100644
index 0000000..5fa87d8
--- /dev/null
+++ b/tests/TestConversationAnalyzer.py
@@ -0,0 +1,231 @@
+import pytest
+from ConversationAnalyzer import ConversationAnalyzer
+from People import People
+from utils import dt
+
+TEST_DATA_PATH = '/home/levente/projects/facebook-data-miner/tests/test_data'
+
+
+# @pytest.mark.parametrize("test_input,expected", [("3+5", 8), ("2+4", 6), ("6*9", 42)])
+# def test_eval(test_input, expected):
+#     assert eval(test_input) == expected
+
+# get\(\'.*\'\)\.
+
+
+@pytest.fixture(scope='session')
+def person(people):
+    def _person(name):
+        return people.individuals[name]
+
+    return _person
+
+
+@pytest.fixture(scope='session')
+def analyze(person):
+    def _analyze(name):
+        individual = person(name)
+        return ConversationAnalyzer(name, individual.messages)
+
+    return _analyze
+
+
+@pytest.fixture(scope='session')
+def statistics(person, analyze):
+    def _stats(name, **kwargs):
+        individual = person(name)
+        analyzer = analyze(name)
+        if 'subject' in kwargs or 'start' in kwargs or 'end' in kwargs:  # and others
+            return analyzer.get_stats(individual.messages, **kwargs)
+        else:
+            return analyzer.stats
+
+    return _stats
+
+
+# TODO extend all functions with all the data
+def test_stats_toke_hal_all(statistics):
+    stats = statistics('Tőke Hal')
+
+    assert stats.msg_count == 5
+    assert stats.unique_msg_count == 4
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 6
+    assert stats.unique_word_count == 4
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 17
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_toke_hal_me(statistics):
+    stats = statistics('Tőke Hal', subject='me')
+
+    assert stats.msg_count == 3
+    assert stats.unique_msg_count == 3
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 4
+    assert stats.unique_word_count == 3
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 12
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_toke_hal_partner(statistics):
+    stats = statistics('Tőke Hal', subject='partner')
+
+    assert stats.msg_count == 2
+    assert stats.unique_msg_count == 2
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 2
+    assert stats.unique_word_count == 2
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 5
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_toke_hal_all_2014_11(statistics):
+    stats = statistics('Tőke Hal', subject='all', start=dt(2014, 11), period='m')
+
+    assert stats.msg_count == 4
+    assert stats.unique_msg_count == 3
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    # assert stats.word_frequency == 0
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_toke_hal_partner_2014_11(statistics):
+    stats = statistics('Tőke Hal', subject='partner', start=dt(2014, 11), period='m')
+    assert stats.char_count == 5
+    assert stats.word_count == 2
+
+
+def test_stats_toke_hal_me_2014_11(statistics):
+    stats = statistics('Tőke Hal', subject='me', start=dt(2014, 11), period='m')
+    assert stats.unique_word_count == 3
+
+
+#
+def test_stats_toke_hal_all_2014_12(statistics):
+    stats = statistics('Tőke Hal', subject='all', start=dt(2014, 12), period='m')
+    assert stats.msg_count == 1
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.unique_word_count == 1
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 3
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_toke_hal_partner_2014_12(statistics):
+    stats = statistics('Tőke Hal', subject='partner', start=dt(2014, 12), period='m')
+    assert stats.word_count == 0
+
+
+def test_stats_toke_hal_me_2014_12(statistics):
+    stats = statistics('Tőke Hal', subject='me', start=dt(2014, 12), period='m')
+    assert stats.unique_msg_count == 1
+
+
+def test_stats_teflon_musk(statistics):
+    stats = statistics('Teflon Musk')
+    assert stats.msg_count == 6
+    assert stats.unique_msg_count == 2  # TODO this does not count media messages
+    # assert stats.most_used_msgs == 0 # TODO should only return the most used or e.g. top10 most used
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 14
+    assert stats.unique_word_count == 7
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 52  # 23
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_teflon_musk_me(statistics):
+    stats = statistics('Teflon Musk', subject='me')
+    assert stats.msg_count == 3
+    assert stats.unique_msg_count == 1
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 12
+    assert stats.unique_word_count == 6
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 48
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_teflon_musk_partner(statistics):
+    stats = statistics('Teflon Musk', subject='partner')
+    assert stats.msg_count == 3
+    assert stats.unique_msg_count == 1
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 2
+    assert stats.unique_word_count == 1
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 4
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_teflon_musk_all_2014_9(statistics):
+    stats = statistics('Teflon Musk', subject='all', start=dt(2014, 9), period='m')
+    assert stats.msg_count == 1
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 6
+    # assert stats.word_frequency == 0
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_teflon_musk_me_2014_9(statistics):
+    stats = statistics('Teflon Musk', subject='me', start=dt(2014, 9), period='m')
+    assert stats.unique_word_count == 6
+
+
+def test_stats_teflon_musk_partner_2014_9(statistics):
+    stats = statistics('Teflon Musk', subject='partner', start=dt(2014, 9), period='m')
+    assert stats.unique_msg_count == 0
+    assert stats.char_count == 0
+
+
+def test_stats_teflon_musk_all_2014_11(statistics):
+    stats = statistics('Teflon Musk', subject='all', start=dt(2014, 11), period='m')
+    assert stats.msg_count == 4
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    # assert stats.word_frequency == 0
+    # assert stats.most_used_chars == 0
+
+
+def test_stats_teflon_musk_me_2014_11(statistics):
+    stats = statistics('Teflon Musk', subject='me', start=dt(2014, 11), period='m')
+    assert stats.word_count == 6
+
+
+def test_stats_teflon_musk_partner_2014_11(statistics):
+    stats = statistics('Teflon Musk', subject='partner', start=dt(2014, 11), period='m')
+    assert stats.unique_msg_count == 1
+    assert stats.unique_word_count == 1
+    assert stats.char_count == 4
+
+
+def test_stats_teflon_musk_all_2014_12(statistics):
+    stats = statistics('Teflon Musk', subject='all', start=dt(2014, 12), period='m')
+
+    assert stats.msg_count == 1
+    assert stats.unique_msg_count == 0
+    # assert stats.most_used_msgs == 0
+    # assert stats.msg_frequency == 0
+    assert stats.word_count == 0
+    assert stats.unique_word_count == 0
+    # assert stats.word_frequency == 0
+    assert stats.char_count == 0
+    # assert stats.most_used_chars == 0
+
+
+def test_time_series_analysis_for_user(analyze):
+    analyzer = analyze('Teflon Musk')
+    analyzer.get_time_series_data(subject='all')
+    assert 1
diff --git a/tests/TestConversations.py b/tests/TestConversations.py
new file mode 100644
index 0000000..e198dc9
--- /dev/null
+++ b/tests/TestConversations.py
@@ -0,0 +1,71 @@
+import pandas as pd
+import pytest
+from Conversations import Conversations
+import os
+TEST_DATA_PATH = '/home/levente/projects/facebook-data-miner/tests/test_data'
+
+
+@pytest.fixture()
+def convos():
+    convo = Conversations(f'{TEST_DATA_PATH}')
+    return convo.get_people()
+
+
+def test_get_all_people_from_convo(convos):
+    people = []
+
+    for convo in convos.keys():
+        if convo.startswith('group'):
+            people += [p for p in convos[convo].get('participants')]
+        else:
+            people.append(convo)
+    people = list(set(people))
+
+    expected = ['Dér Dénes', 'Facebook User', 'Foo Bar', 'John Doe', 'Teflon Musk', 'Benedek Elek', 'Donald Duck',
+                'Tőke Hal']
+    # TODO LATER what to do with Facebook User??????
+    assert sorted(people) == sorted(expected)
+
+
+def test_all_convos_have_dir(convos):
+    assert all([data.get('messages_dir') for data in convos.values()])
+
+
+def test_all_convos_have_messages_df(convos):
+    assert all([isinstance(data.get('messages'), pd.DataFrame) for data in convos.values()])
+
+
+def test_some_convos_as_media_dir(convos):
+    assert convos.get('Teflon Musk').get('media_dir')
+    assert not convos.get('Benedek Elek').get('media_dir')
+
+def test_convo_media_has_one_folder_of_possibles(convos):
+    listed_dir = os.listdir(f"{TEST_DATA_PATH}/{convos.get('Teflon Musk').get('media_dir')}")
+    assert 'files' in listed_dir
+    assert 'photos' in listed_dir
+    assert 'audio' not in listed_dir
+
+def test_groups_have_more_than_two_participates(convos):
+    groups = {convo: data for convo, data in convos.items() if convo.startswith('group')}
+    # TODO participants should contain the user itself as well
+    assert all([len(data.get('participants')) > 2 for data in groups.values()])
+
+
+
+"""
+testcases:
+- individual convos contain all names, compact_names, message folders and media folders
+  - media folders are a big question. how do you get it? actually once you have the thread_path then from that you can guess,
+  OR better off use the uri in the messages... fuck seems complicated
+- friends contain all names and compact names,
+- convos and friends has a common set, and the set is identical
+- people gets assigned with all the unique friends and individual/group convos
+
+gonna test:
+- assigning messages to friends,
+- deal with multiple directories, IF there are multiple directories,
+- 
+concerns:
+- what to do with non-friends,
+- I assume multiple directories are because of files sent,
+"""
diff --git a/tests/TestFriends.py b/tests/TestFriends.py
new file mode 100644
index 0000000..f336609
--- /dev/null
+++ b/tests/TestFriends.py
@@ -0,0 +1,38 @@
+import pytest
+
+from Friends import Friends
+
+TEST_DATA_PATH = '/home/levente/projects/facebook-data-miner/tests/test_data'
+
+
+@pytest.fixture()
+def expected_friends():
+    return {'John Doe': {'compact_name': 'johndoe', 'path': None},
+            'Donald Duck': {'compact_name': 'donaldduck', 'path': None},
+            'Szett Droxler': {'compact_name': 'szettdroxler', 'path': None},
+            'Foo Bar': {'compact_name': 'foobar', 'path': None},
+            'Tőke Hal': {'compact_name': 'tokehal', 'path': None},
+            'Dér Dénes': {'compact_name': 'derdenes', 'path': None},
+            'Teflon Musk': {'compact_name': 'teflonmusk', 'path': None},
+            'Daisy Duck': {'compact_name': 'daisyduck', 'path': None},
+            'Guy Fawkes': {'compact_name': 'guyfawkes', 'path': None}}
+
+
+@pytest.fixture()
+def friends():
+    f = Friends(f'{TEST_DATA_PATH}/friends/friends.json')
+    return f.get_people()
+
+
+def test_get_peoples_names_from_friends(friends, expected_friends):
+    assert all([p in expected_friends.keys() for p in friends])
+
+
+def test_get_peoples_compact_name_from_friends(friends, expected_friends):
+    expected_compact_names = [value.get('compact_name') for value in expected_friends.values()]
+
+    assert all([p.get('compact_name') in expected_compact_names for p in friends.values()])
+
+
+
+# TODO what happens when two friends have same name??
diff --git a/tests/TestMessages.py b/tests/TestMessages.py
new file mode 100644
index 0000000..e69de29
diff --git a/tests/TestMessagingAnalyzer.py b/tests/TestMessagingAnalyzer.py
new file mode 100644
index 0000000..d53af70
--- /dev/null
+++ b/tests/TestMessagingAnalyzer.py
@@ -0,0 +1,143 @@
+import pytest
+from MessagingAnalyzer import MessagingAnalyzer
+from utils import dt
+
+
+@pytest.fixture(scope='session')
+def analyzer(people):
+    return MessagingAnalyzer(people.names, people.individuals)
+
+
+def test_total_number_of_messages(analyzer):
+    assert analyzer.total_number_of_messages() == 14
+    assert analyzer.total_number_of_messages(start=dt(year=2014), period='y') == 11
+    assert analyzer.total_number_of_messages(start=dt(year=2018), period='y') == 3
+
+    assert analyzer.total_number_of_messages(start=dt(year=2014, month=9), period='m') == 1
+    assert analyzer.total_number_of_messages(start=dt(year=2014, month=11), period='m') == 8
+    assert analyzer.total_number_of_messages(start=dt(year=2014, month=12), period='m') == 2
+    assert analyzer.total_number_of_messages(start=dt(year=2018, month=1), period='m') == 3
+
+    assert analyzer.total_number_of_messages(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_messages(start=dt(year=2011, month=11), period='m') == 0
+    assert analyzer.total_number_of_messages(start=dt(year=2018, month=5), period='m') == 0
+
+
+def test_total_number_of_words(analyzer):
+    assert analyzer.total_number_of_words() == 24
+
+    assert analyzer.total_number_of_words(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_words(start=dt(year=2014), period='y') == 20
+    assert analyzer.total_number_of_words(start=dt(year=2018), period='y') == 4
+
+    assert analyzer.total_number_of_words(start=dt(year=2014, month=9), period='m') == 6
+    assert analyzer.total_number_of_words(start=dt(year=2014, month=11), period='m') == 13
+    assert analyzer.total_number_of_words(start=dt(year=2014, month=12), period='m') == 1
+
+    assert analyzer.total_number_of_words(start=dt(year=2018, month=1), period='m') == 4
+    assert analyzer.total_number_of_words(start=dt(year=2018, month=2), period='m') == 0
+
+
+def test_total_number_of_characters(analyzer):
+    assert analyzer.total_number_of_characters() == 81
+
+    assert analyzer.total_number_of_characters(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_characters(start=dt(year=2014), period='y') == 69
+    assert analyzer.total_number_of_characters(start=dt(year=2018), period='y') == 12
+
+    assert analyzer.total_number_of_characters(start=dt(year=2014, month=9), period='m') == 24
+    assert analyzer.total_number_of_characters(start=dt(year=2014, month=11), period='m') == 42
+    assert analyzer.total_number_of_characters(start=dt(year=2014, month=12), period='m') == 3
+
+    assert analyzer.total_number_of_characters(start=dt(year=2018, month=1), period='m') == 12
+    assert analyzer.total_number_of_characters(start=dt(year=2018, month=2), period='m') == 0
+
+
+def test_total_number_of_messages_sent(analyzer):
+    assert analyzer.total_number_of_messages_sent() == 8
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2014), period='y') == 6
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2018), period='y') == 2
+
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2014, month=9), period='m') == 1
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2014, month=11), period='m') == 4
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2014, month=12), period='m') == 1
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2018, month=1), period='m') == 2
+
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2011, month=11), period='m') == 0
+    assert analyzer.total_number_of_messages_sent(start=dt(year=2018, month=5), period='m') == 0
+
+
+def test_total_number_of_words_sent(analyzer):
+    assert analyzer.total_number_of_words_sent() == 19
+
+    assert analyzer.total_number_of_words_sent(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_words_sent(start=dt(year=2014), period='y') == 16
+    assert analyzer.total_number_of_words_sent(start=dt(year=2018), period='y') == 3
+
+    assert analyzer.total_number_of_words_sent(start=dt(year=2014, month=9), period='m') == 6
+    assert analyzer.total_number_of_words_sent(start=dt(year=2014, month=11), period='m') == 9
+    assert analyzer.total_number_of_words_sent(start=dt(year=2014, month=12), period='m') == 1
+
+    assert analyzer.total_number_of_words_sent(start=dt(year=2018, month=1), period='m') == 3
+    assert analyzer.total_number_of_words_sent(start=dt(year=2018, month=2), period='m') == 0
+
+
+def test_total_number_of_characters_sent(analyzer):
+    assert analyzer.total_number_of_characters_sent() == 69
+
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2014), period='y') == 60
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2018), period='y') == 9
+
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2014, month=9), period='m') == 24
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2014, month=11), period='m') == 33
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2014, month=12), period='m') == 3
+
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2018, month=1), period='m') == 9
+    assert analyzer.total_number_of_characters_sent(start=dt(year=2018, month=2), period='m') == 0
+
+
+def test_total_number_of_messages_received(analyzer):
+    assert analyzer.total_number_of_messages_received() == 6
+    assert analyzer.total_number_of_messages_received(start=dt(year=2014), period='y') == 5
+    assert analyzer.total_number_of_messages_received(start=dt(year=2018), period='y') == 1
+
+    assert analyzer.total_number_of_messages_received(start=dt(year=2014, month=9), period='m') == 0
+    assert analyzer.total_number_of_messages_received(start=dt(year=2014, month=11), period='m') == 4
+    assert analyzer.total_number_of_messages_received(start=dt(year=2014, month=12), period='m') == 1
+    assert analyzer.total_number_of_messages_received(start=dt(year=2018, month=1), period='m') == 1
+
+    assert analyzer.total_number_of_messages_received(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_messages_received(start=dt(year=2011, month=11), period='m') == 0
+    assert analyzer.total_number_of_messages_received(start=dt(year=2018, month=5), period='m') == 0
+
+
+def test_total_number_of_words_received(analyzer):
+    assert analyzer.total_number_of_words_received() == 5
+
+    assert analyzer.total_number_of_words_received(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_words_received(start=dt(year=2014), period='y') == 4
+    assert analyzer.total_number_of_words_received(start=dt(year=2018), period='y') == 1
+
+    assert analyzer.total_number_of_words_received(start=dt(year=2014, month=9), period='m') == 0
+    assert analyzer.total_number_of_words_received(start=dt(year=2014, month=11), period='m') == 4
+    assert analyzer.total_number_of_words_received(start=dt(year=2014, month=12), period='m') == 0
+
+    assert analyzer.total_number_of_words_received(start=dt(year=2018, month=1), period='m') == 1
+    assert analyzer.total_number_of_words_received(start=dt(year=2018, month=2), period='m') == 0
+
+
+def test_total_number_of_characters_received(analyzer):
+    assert analyzer.total_number_of_characters_received() == 12
+
+    assert analyzer.total_number_of_characters_received(start=dt(year=2000), period='y') == 0
+    assert analyzer.total_number_of_characters_received(start=dt(year=2014), period='y') == 9
+    assert analyzer.total_number_of_characters_received(start=dt(year=2018), period='y') == 3
+
+    assert analyzer.total_number_of_characters_received(start=dt(year=2014, month=9), period='m') == 0
+    assert analyzer.total_number_of_characters_received(start=dt(year=2014, month=11), period='m') == 9
+    assert analyzer.total_number_of_characters_received(start=dt(year=2014, month=12), period='m') == 0
+
+    assert analyzer.total_number_of_characters_received(start=dt(year=2018, month=1), period='m') == 3
+    assert analyzer.total_number_of_characters_received(start=dt(year=2018, month=2), period='m') == 0
diff --git a/tests/TestPeople.py b/tests/TestPeople.py
new file mode 100644
index 0000000..e29448b
--- /dev/null
+++ b/tests/TestPeople.py
@@ -0,0 +1,36 @@
+import pytest
+
+
+
+@pytest.fixture()
+def people_names():
+    return ['John Doe', 'Donald Duck', 'Szett Droxler', 'Foo Bar', 'Tőke Hal', 'Dér Dénes', 'Teflon Musk', 'Daisy Duck',
+            'Guy Fawkes', 'Benedek Elek']
+
+
+def test_specific_people_has_or_has_not_got_messages(people):
+    # TODO parametrize
+    import pandas as pd
+    assert isinstance(people.data.get('Benedek Elek').get('messages'), pd.DataFrame)
+    assert isinstance(people.data.get('Teflon Musk').get('messages'), pd.DataFrame)
+    assert isinstance(people.data.get('Tőke Hal').get('messages'), pd.DataFrame)
+    assert not isinstance(people.data.get('John Doe').get('messages'), pd.DataFrame)
+    assert not isinstance(people.data.get('Szett Droxler').get('messages'), pd.DataFrame)
+    assert not isinstance(people.data.get('Daisy Duck').get('messages'), pd.DataFrame)
+    assert not isinstance(people.data.get('Guy Fawkes').get('messages'), pd.DataFrame)
+
+
+def test_people_name(people, people_names):
+    people_without_groups = [p for p in people.data.keys() if not p.startswith('group')]
+    assert sorted(people_names) == sorted(people_without_groups)
+
+
+def test_some_convos_are_with_friends(people):
+    assert people.data.get('Teflon Musk').get('friend')
+    assert not people.data.get('Benedek Elek').get('friend')
+
+
+def test_specific_people_has_or_has_not_got_media(people):
+    assert people.data.get('Teflon Musk').get('media_dir')
+
+#TODO test individuals too
\ No newline at end of file
diff --git a/tests/conftest.py b/tests/conftest.py
new file mode 100644
index 0000000..430e923
--- /dev/null
+++ b/tests/conftest.py
@@ -0,0 +1,10 @@
+import pytest
+from People import People
+
+TEST_DATA_PATH = '/home/levente/projects/facebook-data-miner/tests/test_data'
+
+
+@pytest.fixture(scope='session')
+def people():
+    p = People(path=TEST_DATA_PATH)
+    return p
diff --git a/tests/playground.py b/tests/playground.py
new file mode 100644
index 0000000..681c3cb
--- /dev/null
+++ b/tests/playground.py
@@ -0,0 +1,29 @@
+import pytest
+from pytest_cases import parametrize, fixture, fixture_ref, lazy_value
+# @pytest.mark.parametrize("test_input,expected", [("3+5", 8), ("2+4", 6), ("6*9", 42)])
+# def test_eval(test_input, expected):
+#     assert eval(test_input) == expected
+
+@pytest.fixture
+def world_str():
+    return 'world'
+
+
+def whatfun():
+    return 'what'
+
+
+@fixture
+@parametrize('who', [fixture_ref(world_str),
+                     'you'])
+def greetings(who):
+    return 'hello ' + who
+
+
+@parametrize('main_msg', ['nothing',
+                          fixture_ref(world_str),
+                          lazy_value(whatfun),
+                          fixture_ref(greetings)])
+@pytest.mark.parametrize('ending', ['?', '!'])
+def test_prints(main_msg, ending):
+    print(main_msg + ending)
diff --git a/tests/test_data/friends/friends.json b/tests/test_data/friends/friends.json
new file mode 100644
index 0000000..50c6b8f
--- /dev/null
+++ b/tests/test_data/friends/friends.json
@@ -0,0 +1,40 @@
+{
+  "friends": [
+    {
+      "name": "John Doe",
+      "timestamp": 1590673319
+    },
+    {
+      "name": "Donald Duck",
+      "timestamp": 1586461325
+    },
+    {
+      "name": "Szett Droxler",
+      "timestamp": 1584299908
+    },
+    {
+      "name": "Foo Bar",
+      "timestamp": 1584219292
+    },
+    {
+      "name": "T\u00c5\u0091ke Hal",
+      "timestamp": 1582290479
+    },
+    {
+      "name": "D\u00c3\u00a9r D\u00c3\u00a9nes",
+      "timestamp": 1581523312
+    },
+    {
+      "name": "Teflon Musk",
+      "timestamp": 1580999341
+    },
+    {
+      "name": "Daisy Duck",
+      "timestamp": 1580999200
+    },
+    {
+      "name": "Guy Fawkes",
+      "timestamp": 1580999100
+    }
+  ]
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/TeflonMusk_fSD454F/files/1810.04805.pdf b/tests/test_data/messages/inbox/TeflonMusk_fSD454F/files/1810.04805.pdf
new file mode 100644
index 0000000..2394716
Binary files /dev/null and b/tests/test_data/messages/inbox/TeflonMusk_fSD454F/files/1810.04805.pdf differ
diff --git a/tests/test_data/messages/inbox/TeflonMusk_fSD454F/photos/index.jpeg b/tests/test_data/messages/inbox/TeflonMusk_fSD454F/photos/index.jpeg
new file mode 100644
index 0000000..57fae9c
Binary files /dev/null and b/tests/test_data/messages/inbox/TeflonMusk_fSD454F/photos/index.jpeg differ
diff --git a/tests/test_data/messages/inbox/benedekelek_s4f65sdg/message_1.json b/tests/test_data/messages/inbox/benedekelek_s4f65sdg/message_1.json
new file mode 100644
index 0000000..61c06a1
--- /dev/null
+++ b/tests/test_data/messages/inbox/benedekelek_s4f65sdg/message_1.json
@@ -0,0 +1,34 @@
+{
+  "participants": [
+    {
+      "name": "Benedek Elek"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1515619045145,
+      "content": "not much",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Benedek Elek",
+      "timestamp_ms": 1515618506047,
+      "content": "zup",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1515571228715,
+      "content": "yo",
+      "type": "Generic"
+    }
+  ],
+  "title": "Benedek Elek",
+  "is_still_participant": true,
+  "thread_type": "Regular",
+  "thread_path": "inbox/benedekelek_s4f65sdg"
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/marathon_sfFSFiD76/gifs/21297336_10214236646047101_8870179296803553280_n_2116299178397271.gif b/tests/test_data/messages/inbox/marathon_sfFSFiD76/gifs/21297336_10214236646047101_8870179296803553280_n_2116299178397271.gif
new file mode 100644
index 0000000..666b745
Binary files /dev/null and b/tests/test_data/messages/inbox/marathon_sfFSFiD76/gifs/21297336_10214236646047101_8870179296803553280_n_2116299178397271.gif differ
diff --git a/tests/test_data/messages/inbox/marathon_sfFSFiD76/photos/index.jpeg b/tests/test_data/messages/inbox/marathon_sfFSFiD76/photos/index.jpeg
new file mode 100644
index 0000000..57fae9c
Binary files /dev/null and b/tests/test_data/messages/inbox/marathon_sfFSFiD76/photos/index.jpeg differ
diff --git a/tests/test_data/messages/inbox/marathon_sffsfid76/message_1.json b/tests/test_data/messages/inbox/marathon_sffsfid76/message_1.json
new file mode 100644
index 0000000..89ba754
--- /dev/null
+++ b/tests/test_data/messages/inbox/marathon_sffsfid76/message_1.json
@@ -0,0 +1,85 @@
+{
+  "participants": [
+    {
+      "name": "Teflon Musk"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    },
+    {
+      "name": "Foo Bar"
+    },
+    {
+      "name": "Donald Duck"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1524142359709,
+      "content": ":D",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1524137912776,
+      "content": "hmmm",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1524137882444,
+      "content": "we could go but running is free",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1524137859673,
+      "photos": [
+        {
+          "uri": "messages/inbox/marathon_sfFSFiD76/photos/index.jpeg",
+          "creation_timestamp": 1524137857
+        }
+      ],
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1524137749717,
+      "content": "i start today",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1524137737066,
+      "gifs": [
+        {
+          "uri": "messages/inbox/marathon_sfFSFiD76/gifs/21297336_10214236646047101_8870179296803553280_n_2116299178397271.gif"
+        }
+      ],
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1524133955273,
+      "content": "You named the group marathon.",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Foo Bar",
+      "timestamp_ms": 1524133941074,
+      "content": "yapp yapp :D",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1524133902152,
+      "content": "marathon?",
+      "type": "Generic"
+    }
+  ],
+  "title": "marathon",
+  "is_still_participant": true,
+  "thread_type": "RegularGroup",
+  "thread_path": "inbox/marathon_sffsfid76"
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/teflonmusk_fsd454f/message_1.json b/tests/test_data/messages/inbox/teflonmusk_fsd454f/message_1.json
new file mode 100644
index 0000000..1c35b5e
--- /dev/null
+++ b/tests/test_data/messages/inbox/teflonmusk_fsd454f/message_1.json
@@ -0,0 +1,60 @@
+{
+  "participants": [
+    {
+      "name": "Teflon Musk"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "Teflon Musk",
+      "timestamp_ms": 1419620506047,
+      "files": [
+        {
+          "uri": "messages/inbox/TeflonMusk_fSD454F/files/1810.04805.pdf"
+        }
+      ],
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1416619045145,
+      "photos": [
+        {
+          "uri": "messages/inbox/TeflonMusk_fSD454F/photos/index.jpeg"
+        }
+      ],
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Teflon Musk",
+      "timestamp_ms": 1415618806047,
+      "content": "no",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Teflon Musk",
+      "timestamp_ms": 1415618506047,
+      "content": "no",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1415571228715,
+      "content": "are you the real teflon musk?",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1411570928715,
+      "content": "are you the real teflon musk?",
+      "type": "Generic"
+    }
+  ],
+  "title": "Teflon Musk",
+  "is_still_participant": true,
+  "thread_type": "Regular",
+  "thread_path": "inbox/teflonmusk_fsd454f"
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/teflonmuskfoobarandjohndoe_sf5d4f56sd_gf/message_1.json b/tests/test_data/messages/inbox/teflonmuskfoobarandjohndoe_sf5d4f56sd_gf/message_1.json
new file mode 100644
index 0000000..84da80b
--- /dev/null
+++ b/tests/test_data/messages/inbox/teflonmuskfoobarandjohndoe_sf5d4f56sd_gf/message_1.json
@@ -0,0 +1,46 @@
+{
+  "participants": [
+    {
+      "name": "Teflon Musk"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    },
+    {
+      "name": "John Doe"
+    },
+    {
+      "name": "Foo Bar"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "John Doe",
+      "timestamp_ms": 1310907774237,
+      "content": "ok",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Teflon Musk",
+      "timestamp_ms": 1310907632012,
+      "content": "basic group messages",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Foo Bar",
+      "timestamp_ms": 1310907613721,
+      "content": "what do you test",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1310907608580,
+      "content": "test",
+      "type": "Generic"
+    }
+  ],
+  "title": "Foo Bar, John Doe and Teflon Musk",
+  "is_still_participant": true,
+  "thread_type": "RegularGroup",
+  "thread_path": "inbox/teflonmuskfoobarandjohndoe_sf5d4f56sd_gf"
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/tokehal_sdf7fs9d876/message_1.json b/tests/test_data/messages/inbox/tokehal_sdf7fs9d876/message_1.json
new file mode 100644
index 0000000..66fc886
--- /dev/null
+++ b/tests/test_data/messages/inbox/tokehal_sdf7fs9d876/message_1.json
@@ -0,0 +1,46 @@
+{
+  "participants": [
+    {
+      "name": "T\u00c5\u0091ke Hal"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "Levente CsÅ\u0091ke",
+      "timestamp_ms": 1417619245145,
+      "content": "not",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1415619045145,
+      "content": "not much",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "T\u00c5\u0091ke Hal",
+      "timestamp_ms": 1415618506047,
+      "content": "zup",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "T\u00c5\u0091ke Hal",
+      "timestamp_ms": 1415618406047,
+      "content": "yo",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Levente Cs\u00c5\u0091ke",
+      "timestamp_ms": 1415571228715,
+      "content": "yo",
+      "type": "Generic"
+    }
+  ],
+  "title": "T\u00c5\u0091ke Hal",
+  "is_still_participant": true,
+  "thread_type": "Regular",
+  "thread_path": "inbox/tokehal_sdf7fs9d876"
+}
\ No newline at end of file
diff --git a/tests/test_data/messages/inbox/tokehalfoobardonaldduckand2others_safsdf_fdf3/message_1.json b/tests/test_data/messages/inbox/tokehalfoobardonaldduckand2others_safsdf_fdf3/message_1.json
new file mode 100644
index 0000000..d9dfd9a
--- /dev/null
+++ b/tests/test_data/messages/inbox/tokehalfoobardonaldduckand2others_safsdf_fdf3/message_1.json
@@ -0,0 +1,58 @@
+{
+  "participants": [
+    {
+      "name": "T\u00c5\u0091ke Hal"
+    },
+    {
+      "name": "Levente Cs\u00c5\u0091ke"
+    },
+    {
+      "name": "D\u00c3\u00a9r D\u00c3\u00a9nes"
+    },
+    {
+      "name": "Facebook User"
+    },
+    {
+      "name": "Donald Duck"
+    },
+    {
+      "name": "Foo Bar"
+    }
+  ],
+  "messages": [
+    {
+      "sender_name": "Facebook User",
+      "timestamp_ms": 1310907774237,
+      "content": "ok",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "D\u00c3\u00a9r D\u00c3\u00a9nes",
+      "timestamp_ms": 1310907632012,
+      "content": "blabla",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "T\u00c5\u0091ke Hal",
+      "timestamp_ms": 1310907632012,
+      "content": "basic group messages",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Foo Bar",
+      "timestamp_ms": 1310907613721,
+      "content": "what do you test",
+      "type": "Generic"
+    },
+    {
+      "sender_name": "Donald Duck",
+      "timestamp_ms": 1310907608580,
+      "content": "test",
+      "type": "Generic"
+    }
+  ],
+  "title": "T\u00c5\u0091ke Hal, Foo Bar, Donald Duck and 2 others",
+  "is_still_participant": true,
+  "thread_type": "RegularGroup",
+  "thread_path": "inbox/tokehalfoobardonaldduckand2others_safsdf_fdf3"
+}
\ No newline at end of file
diff --git a/tests/test_utils.py b/tests/test_utils.py
new file mode 100644
index 0000000..76352c0
--- /dev/null
+++ b/tests/test_utils.py
@@ -0,0 +1,58 @@
+import unittest
+from utils import *
+from pathlib import Path
+import reusables
+from reusables.cli import *
+
+
+#TODO convert this to pytest
+
+class TestUtils(unittest.TestCase):
+    def setUp(self):
+        self.test_data_path = Path(f'{pwd()}/test_data')
+
+    def test_read_json(self):
+        dummy1 = {'0': 'a', '2': 'c'}
+        dummy1_path = Path(self.test_data_path / 'dummy1.json')
+        dump_to_json(file=dummy1_path, data=dummy1)
+
+        dummy2 = {'1': 'b', '3': 'd'}
+        dummy2_path = Path(self.test_data_path / 'dummy2.json')
+        dump_to_json(file=dummy2_path, data=dummy2)
+
+        dummy_unified = get_messages(dummy1_path, dummy2_path)
+        expected = {'0': 'a', '1': 'b', '2': 'c', '3': 'd'}
+
+        self.assertDictEqual(expected, dummy_unified)
+
+        dummy1_path.unlink()
+        dummy2_path.unlink()
+
+    def test_decode_text(self):
+        dummy = {'0': '\u00c5\u0091', '1': '\u00c3\u00a1', '2': ['\u00c5\u0091', {'0': '\u00c3\u00a1'}],
+                 '3': {'0': '\u00c5\u0091'}}
+        dummy_path = Path(self.test_data_path / 'dummy.json')
+        dump_to_json(file=dummy_path, data=dummy)
+        read = get_messages(dummy_path, decode=True)
+        expected = {'0': 'ő', '1': 'á', '2': ['ő', {'0': 'á'}],
+                    '3': {'0': 'ő'}}
+        self.assertEqual(expected, read)
+
+        dummy_path.unlink()
+
+    def later_tests(self):
+        data = {'participants': [{'name': 'Csőke Boglárka'}, {'name': 'Levente Csőke'}],
+                'title': 'Csőke Boglárka',
+                'is_still_participant': True,
+                'thread_type': 'Regular',
+                'thread_path': 'inbox/CsokeBoglarka_5A48Zi9P1w',
+                'messages': [...]}
+        msg_element = {'sender_name': 'Levente Csőke',
+                       'timestamp_ms': 1440948801592,
+                       'content': 'ahaa',
+                       'type': 'Generic'}
+
+    def test_read_real_data(self):
+        bogi_msg = Path('/home/levente/projects/facebook-data-miner/data/messages/inbox/csokeboglarka_5a48zi9p1w')
+        data = get_messages(bogi_msg/'message_1.json', bogi_msg/'message_2.json')
+        msg = data['messages']
\ No newline at end of file
diff --git a/utils.py b/utils.py
new file mode 100644
index 0000000..8594258
--- /dev/null
+++ b/utils.py
@@ -0,0 +1,191 @@
+import json
+import pandas as pd
+import dateparser
+from datetime import datetime, timedelta
+from dateutil.relativedelta import relativedelta
+
+MONTHS = ['january', 'february', 'march', 'april', 'may', 'june', 'july', 'august', 'september', 'october',
+          'november', 'december']
+
+
+def get_stats_for_intervals(func, df, time_series, subject='all'):
+    data = {}
+    for offset, series in time_series.items():
+        data[offset] = {}
+        for i in range(len(series) - 1):  # only looping len - 1 times
+            start = series[i]
+            # TODO will we miss the last entry? I dont think so (99%), but check and correct hand in hand with the timeseries bug
+            # IT DOES NOT! HOWEVER test it with new data injected/modified at runtime <- this is hard
+            end = series[i + 1]
+            data[offset][start] = func(df, subject=subject, start=start, end=end)
+    return data
+
+
+# @date_checker
+def dt(year: int = 2004, month: int = 1, day: int = 1, hour: int = 0):
+    return datetime(year=year, month=month, day=day, hour=hour)
+
+
+def get_messages(*files, decode=True):
+    data = {}
+    for file in files:
+        temp = decode_text(read_json(file)) if decode else read_json(file)
+        if not data:
+            data = temp
+        elif data.get('messages') and temp.get('messages'):
+            data['messages'] += temp.get('messages')
+            if sorted(temp.keys()) != sorted(data.keys()):
+                data = {**temp, **data}
+    return data
+
+
+def read_json(file):
+    with open(file) as f:
+        return json.load(f)
+
+
+def dump_to_json(data=None, file=None):
+    with open(file, 'w') as f:
+        json.dump(data, f)
+
+
+def decode_text(obj):
+    if isinstance(obj, str):
+        return obj.encode('latin_1').decode('utf-8')
+
+    if isinstance(obj, list):
+        return [decode_text(o) for o in obj]
+
+    if isinstance(obj, dict):
+        return {key: decode_text(item) for key, item in obj.items()}
+
+    return obj
+
+
+def order_list_of_dicts(lst, key='timestamp_ms'):
+    return sorted(lst, key=lambda k: k[key])
+
+
+accents_map = {
+    "á": "a",
+    "é": "e",
+    "í": "i",
+    "ó": "o",
+    "ö": "o",
+    "ő": "o",
+    "ú": "u",
+    "ü": "u",
+    "ű": "u",
+    # "Á": "A",
+    # "É": "E",
+    # "Í": "I",
+    # "Ó": "O",
+    # "Ö": "O",
+    # "Ő": "O",
+    # "Ú": "U",
+    # "Ü": "U",
+    # "Ű": "U",
+}
+
+
+#
+
+
+def year_converter(func):
+    """
+    Higher-order function that converts @year param passed to @func into numeric version.
+    @param func:
+    @return:
+    """
+
+    def wrapper(*args, **kwargs):
+        if not kwargs.get('year'):
+            return func(*args, **kwargs)
+        if not isinstance(kwargs.get('year'), int):
+            if kwargs.get('year').isdigit():
+                kwargs['year'] = int(kwargs.get('year'))
+            else:
+                print(f'Year is not a digit. Given year: {kwargs.get("year")}')
+        return func(*args, **kwargs)
+
+    return wrapper
+
+
+def month_converter(func):
+    """
+    Higher-order function that converts @month param passed to @func into numeric version.
+    @param func:
+    @return:
+    """
+
+    def wrapper(*args, **kwargs):
+        if not kwargs.get('month'):
+            return func(*args, **kwargs)
+        if isinstance(kwargs['month'], str) and not kwargs['month'].isdigit():
+            kwargs['month'] = MONTHS.index(kwargs['month'].lower()) + 1
+        return func(*args, **kwargs)
+
+    return wrapper
+
+
+# TODO period can refer to 2 things
+# 1. I used it for y/m/d/h
+# 2. return only with one date_range series
+# decide wtf
+def generate_time_series(start=None, end=None, period=None):
+    start = start or datetime(year=2009, month=10, day=2, hour=0)
+    end = end or datetime.now()
+    time_series = {
+        'y': pd.date_range(start=start, end=end, freq='YS'),  # TODO does not include 2009
+        'm': pd.date_range(start=start, end=end, freq='1MS'),  # TODO does not include october
+        'd': pd.date_range(start=start, end=end, freq='1D'),  # TODO does not include 2. ?! not sure if it is true
+        # TODO put this back after dev phase is over
+        # 'h': pd.date_range(start=start, end=end, freq='1H'), # TODO hour should only be run ONCE
+    }
+    if period and period in ('y', 'm', 'd', 'h'):
+        return time_series[period]
+    return time_series
+
+
+def subject_checker(func):
+    def wrapper(*args, **kwargs):
+        if not kwargs.get('subject') or kwargs.get('subject') not in ('all', 'me', 'partner'):
+            raise ValueError('Parameter `subject` should be one of {all, me, partner}')
+        return func(*args, **kwargs)
+
+    return wrapper
+
+
+def date_checker(func):
+    def wrapper(*args, **kwargs):
+        if kwargs.get('start') is not None and isinstance(kwargs.get('start'), str):
+            kwargs['start'] = dateparser.parse(kwargs.get('start'))
+        if kwargs.get('end') is not None and isinstance(kwargs.get('end'), str):
+            kwargs['end'] = dateparser.parse(kwargs.get('end'))
+        if kwargs.get('start') is None and kwargs.get('end') is None:
+            kwargs['start'] = datetime(year=2006, month=1, day=1)  # foundation date of Facebook
+            kwargs['end'] = datetime.now()
+        if sum(map(bool, [kwargs.get('start'), kwargs.get('end'), kwargs.get('period')])) < 2:
+            raise ValueError(
+                'At least two of the following three input variables has to be passed: {start, end, period}')
+        return func(*args, **kwargs)
+
+    return wrapper
+
+
+def period_checker(func):
+    def wrapper(*args, **kwargs):
+        if kwargs.get('start') is not None and kwargs.get('end') is not None:
+            return func(*args, **kwargs)
+        delta_map = {
+            'y': relativedelta(years=+1),
+            'm': relativedelta(months=+1),
+            'd': timedelta(days=1),
+            'h': timedelta(hours=1)
+        }
+        if not kwargs.get('period') or delta_map[kwargs.get('period')] is None:
+            raise ValueError('Parameter `period` should be one of {y, m, d, h}')
+        kwargs['period'] = delta_map[kwargs.get('period')]
+        return func(*args, **kwargs)
+
+    return wrapper