Fix use of category parameter for search() function

nominatim/api/search/db_search_builder.py

@@ -89,29 +89,31 @@ def build(self, assignment: TokenAssignment) -> Iterator[dbs.AbstractSearch]:
         if sdata is None:
             return
 
-        categories = self.get_search_categories(assignment)
+        near_items = self.get_near_items(assignment)
+        if near_items is not None and not near_items:
+            return # impossible compbination of near items and category parameter
 
         if assignment.name is None:
-            if categories and not sdata.postcodes:
-                sdata.qualifiers = categories
-                categories = None
+            if near_items and not sdata.postcodes:
+                sdata.qualifiers = near_items
+                near_items = None
                 builder = self.build_poi_search(sdata)
             elif assignment.housenumber:
                 hnr_tokens = self.query.get_tokens(assignment.housenumber,
                                                    TokenType.HOUSENUMBER)
                 builder = self.build_housenumber_search(sdata, hnr_tokens, assignment.address)
             else:
                 builder = self.build_special_search(sdata, assignment.address,
-                                                    bool(categories))
+                                                    bool(near_items))
         else:
             builder = self.build_name_search(sdata, assignment.name, assignment.address,
-                                             bool(categories))
+                                             bool(near_items))
 
-        if categories:
-            penalty = min(categories.penalties)
-            categories.penalties = [p - penalty for p in categories.penalties]
+        if near_items:
+            penalty = min(near_items.penalties)
+            near_items.penalties = [p - penalty for p in near_items.penalties]
             for search in builder:
-                yield dbs.NearSearch(penalty + assignment.penalty, categories, search)
+                yield dbs.NearSearch(penalty + assignment.penalty, near_items, search)
         else:
             for search in builder:
                 search.penalty += assignment.penalty
@@ -321,8 +323,15 @@ def get_search_data(self, assignment: TokenAssignment) -> Optional[dbf.SearchDat
                               self.query.get_tokens(assignment.postcode,
                                                     TokenType.POSTCODE))
         if assignment.qualifier:
-            sdata.set_qualifiers(self.query.get_tokens(assignment.qualifier,
-                                                       TokenType.QUALIFIER))
+            tokens = self.query.get_tokens(assignment.qualifier, TokenType.QUALIFIER)
+            if self.details.categories:
+                tokens = [t for t in tokens if t.get_category() in self.details.categories]
+                if not tokens:
+                    return None
+            sdata.set_qualifiers(tokens)
+        elif self.details.categories:
+            sdata.qualifiers = dbf.WeightedCategories(self.details.categories,
+                                                      [0.0] * len(self.details.categories))
 
         if assignment.address:
             sdata.set_ranking([self.get_addr_ranking(r) for r in assignment.address])
@@ -332,25 +341,23 @@ def get_search_data(self, assignment: TokenAssignment) -> Optional[dbf.SearchDat
         return sdata
 
 
-    def get_search_categories(self,
-                              assignment: TokenAssignment) -> Optional[dbf.WeightedCategories]:
-        """ Collect tokens for category search or use the categories
+    def get_near_items(self, assignment: TokenAssignment) -> Optional[dbf.WeightedCategories]:
+        """ Collect tokens for near items search or use the categories
             requested per parameter.
             Returns None if no category search is requested.
         """
-        if assignment.category:
+        if assignment.near_item:
             tokens: Dict[Tuple[str, str], float] = {}
-            for t in self.query.get_tokens(assignment.category, TokenType.CATEGORY):
+            for t in self.query.get_tokens(assignment.near_item, TokenType.NEAR_ITEM):
                 cat = t.get_category()
+                # The category of a near search will be that of near_item.
+                # Thus, if search is restricted to a category parameter,
+                # the two sets must intersect.
                 if (not self.details.categories or cat in self.details.categories)\
                    and t.penalty < tokens.get(cat, 1000.0):
                     tokens[cat] = t.penalty
             return dbf.WeightedCategories(list(tokens.keys()), list(tokens.values()))
 
-        if self.details.categories:
-            return dbf.WeightedCategories(self.details.categories,
-                                          [0.0] * len(self.details.categories))
-
         return None
 
 

nominatim/api/search/db_searches.py

@@ -766,9 +766,6 @@ async def lookup(self, conn: SearchConnection,
             assert result
             result.bbox = Bbox.from_wkb(row.bbox)
             result.accuracy = row.accuracy
-            if not details.excluded or not result.place_id in details.excluded:
-                results.append(result)
-
             if self.housenumbers and row.rank_address < 30:
                 if row.placex_hnr:
                     subs = _get_placex_housenumbers(conn, row.placex_hnr, details)
@@ -788,6 +785,14 @@ async def lookup(self, conn: SearchConnection,
                             sub.accuracy += 0.6
                         results.append(sub)
 
-                result.accuracy += 1.0 # penalty for missing housenumber
+                # Only add the street as a result, if it meets all other
+                # filter conditions.
+                if (not details.excluded or result.place_id not in details.excluded)\
+                   and (not self.qualifiers or result.category in self.qualifiers.values)\
+                   and result.rank_address >= details.min_rank:
+                    result.accuracy += 1.0 # penalty for missing housenumber
+                    results.append(result)
+            else:
+                results.append(result)
 
         return results

nominatim/api/search/icu_tokenizer.py

@@ -184,13 +184,13 @@ async def analyze_query(self, phrases: List[qmod.Phrase]) -> qmod.QueryStruct:
                 if row.type == 'S':
                     if row.info['op'] in ('in', 'near'):
                         if trange.start == 0:
-                            query.add_token(trange, qmod.TokenType.CATEGORY, token)
+                            query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
                     else:
                         query.add_token(trange, qmod.TokenType.QUALIFIER, token)
                         if trange.start == 0 or trange.end == query.num_token_slots():
                             token = copy(token)
                             token.penalty += 0.1 * (query.num_token_slots())
-                            query.add_token(trange, qmod.TokenType.CATEGORY, token)
+                            query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
                 else:
                     query.add_token(trange, DB_TO_TOKEN_TYPE[row.type], token)
 

nominatim/api/search/legacy_tokenizer.py

@@ -107,15 +107,15 @@ async def analyze_query(self, phrases: List[qmod.Phrase]) -> qmod.QueryStruct:
         for row in await self.lookup_in_db(lookup_words):
             for trange in words[row.word_token.strip()]:
                 token, ttype = self.make_token(row)
-                if ttype == qmod.TokenType.CATEGORY:
+                if ttype == qmod.TokenType.NEAR_ITEM:
                     if trange.start == 0:
-                        query.add_token(trange, qmod.TokenType.CATEGORY, token)
+                        query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
                 elif ttype == qmod.TokenType.QUALIFIER:
                     query.add_token(trange, qmod.TokenType.QUALIFIER, token)
                     if trange.start == 0 or trange.end == query.num_token_slots():
                         token = copy(token)
                         token.penalty += 0.1 * (query.num_token_slots())
-                        query.add_token(trange, qmod.TokenType.CATEGORY, token)
+                        query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
                 elif ttype != qmod.TokenType.PARTIAL or trange.start + 1 == trange.end:
                     query.add_token(trange, ttype, token)
 
@@ -195,7 +195,7 @@ def make_token(self, row: SaRow) -> Tuple[LegacyToken, qmod.TokenType]:
                 ttype = qmod.TokenType.POSTCODE
                 lookup_word = row.word_token[1:]
             else:
-                ttype = qmod.TokenType.CATEGORY if row.operator in ('in', 'near')\
+                ttype = qmod.TokenType.NEAR_ITEM if row.operator in ('in', 'near')\
                         else qmod.TokenType.QUALIFIER
                 lookup_word = row.word
         elif row.word_token.startswith(' '):

nominatim/api/search/query.py

@@ -46,7 +46,7 @@ class TokenType(enum.Enum):
     """ Country name or reference. """
     QUALIFIER = enum.auto()
     """ Special term used together with name (e.g. _Hotel_ Bellevue). """
-    CATEGORY = enum.auto()
+    NEAR_ITEM = enum.auto()
     """ Special term used as searchable object(e.g. supermarket in ...). """
 
 
@@ -78,7 +78,7 @@ def compatible_with(self, ttype: TokenType,
             return not is_full_phrase or ttype != TokenType.QUALIFIER
         if self == PhraseType.AMENITY:
             return ttype in (TokenType.WORD, TokenType.PARTIAL)\
-                   or (is_full_phrase and ttype == TokenType.CATEGORY)\
+                   or (is_full_phrase and ttype == TokenType.NEAR_ITEM)\
                    or (not is_full_phrase and ttype == TokenType.QUALIFIER)
         if self == PhraseType.STREET:
             return ttype in (TokenType.WORD, TokenType.PARTIAL, TokenType.HOUSENUMBER)

nominatim/api/search/token_assignment.py

@@ -46,7 +46,7 @@ class TokenAssignment: # pylint: disable=too-many-instance-attributes
     housenumber: Optional[qmod.TokenRange] = None
     postcode: Optional[qmod.TokenRange] = None
     country: Optional[qmod.TokenRange] = None
-    category: Optional[qmod.TokenRange] = None
+    near_item: Optional[qmod.TokenRange] = None
     qualifier: Optional[qmod.TokenRange] = None
 
 
@@ -64,8 +64,8 @@ def from_ranges(ranges: TypedRangeSeq) -> 'TokenAssignment':
                 out.postcode = token.trange
             elif token.ttype == qmod.TokenType.COUNTRY:
                 out.country = token.trange
-            elif token.ttype == qmod.TokenType.CATEGORY:
-                out.category = token.trange
+            elif token.ttype == qmod.TokenType.NEAR_ITEM:
+                out.near_item = token.trange
             elif token.ttype == qmod.TokenType.QUALIFIER:
                 out.qualifier = token.trange
         return out
@@ -109,7 +109,7 @@ def is_final(self) -> bool:
         """
         # Country and category must be the final term for left-to-right
         return len(self.seq) > 1 and \
-               self.seq[-1].ttype in (qmod.TokenType.COUNTRY, qmod.TokenType.CATEGORY)
+               self.seq[-1].ttype in (qmod.TokenType.COUNTRY, qmod.TokenType.NEAR_ITEM)
 
 
     def appendable(self, ttype: qmod.TokenType) -> Optional[int]:
@@ -165,22 +165,22 @@ def appendable(self, ttype: qmod.TokenType) -> Optional[int]:
         if ttype == qmod.TokenType.COUNTRY:
             return None if self.direction == -1 else 1
 
-        if ttype == qmod.TokenType.CATEGORY:
+        if ttype == qmod.TokenType.NEAR_ITEM:
             return self.direction
 
         if ttype == qmod.TokenType.QUALIFIER:
             if self.direction == 1:
                 if (len(self.seq) == 1
-                    and self.seq[0].ttype in (qmod.TokenType.PARTIAL, qmod.TokenType.CATEGORY)) \
+                    and self.seq[0].ttype in (qmod.TokenType.PARTIAL, qmod.TokenType.NEAR_ITEM)) \
                    or (len(self.seq) == 2
-                       and self.seq[0].ttype == qmod.TokenType.CATEGORY
+                       and self.seq[0].ttype == qmod.TokenType.NEAR_ITEM
                        and self.seq[1].ttype == qmod.TokenType.PARTIAL):
                     return 1
                 return None
             if self.direction == -1:
                 return -1
 
-            tempseq = self.seq[1:] if self.seq[0].ttype == qmod.TokenType.CATEGORY else self.seq
+            tempseq = self.seq[1:] if self.seq[0].ttype == qmod.TokenType.NEAR_ITEM else self.seq
             if len(tempseq) == 0:
                 return 1
             if len(tempseq) == 1 and self.seq[0].ttype == qmod.TokenType.HOUSENUMBER:
@@ -253,7 +253,7 @@ def recheck_sequence(self) -> bool:
                 priors = sum(1 for t in self.seq[hnrpos+1:] if t.ttype == qmod.TokenType.PARTIAL)
                 if not self._adapt_penalty_from_priors(priors, 1):
                     return False
-            if any(t.ttype == qmod.TokenType.CATEGORY for t in self.seq):
+            if any(t.ttype == qmod.TokenType.NEAR_ITEM for t in self.seq):
                 self.penalty += 1.0
 
         return True
@@ -368,7 +368,7 @@ def get_assignments(self, query: qmod.QueryStruct) -> Iterator[TokenAssignment]:
 
         # Postcode or country-only search
         if not base.address:
-            if not base.housenumber and (base.postcode or base.country or base.category):
+            if not base.housenumber and (base.postcode or base.country or base.near_item):
                 log().comment('postcode/country search')
                 yield dataclasses.replace(base, penalty=self.penalty)
         else:

test/python/api/search/test_api_search_query.py

@@ -106,11 +106,11 @@ def test_query_struct_amenity_single_word():
     q.add_node(query.BreakType.END, query.PhraseType.NONE)
 
     q.add_token(query.TokenRange(0, 1), query.TokenType.PARTIAL, mktoken(1))
-    q.add_token(query.TokenRange(0, 1), query.TokenType.CATEGORY, mktoken(2))
+    q.add_token(query.TokenRange(0, 1), query.TokenType.NEAR_ITEM, mktoken(2))
     q.add_token(query.TokenRange(0, 1), query.TokenType.QUALIFIER, mktoken(3))
 
     assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.PARTIAL)) == 1
-    assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.CATEGORY)) == 1
+    assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.NEAR_ITEM)) == 1
     assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.QUALIFIER)) == 0
 
 
@@ -121,14 +121,14 @@ def test_query_struct_amenity_two_words():
 
     for trange in [(0, 1), (1, 2)]:
         q.add_token(query.TokenRange(*trange), query.TokenType.PARTIAL, mktoken(1))
-        q.add_token(query.TokenRange(*trange), query.TokenType.CATEGORY, mktoken(2))
+        q.add_token(query.TokenRange(*trange), query.TokenType.NEAR_ITEM, mktoken(2))
         q.add_token(query.TokenRange(*trange), query.TokenType.QUALIFIER, mktoken(3))
 
     assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.PARTIAL)) == 1
-    assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.CATEGORY)) == 0
+    assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.NEAR_ITEM)) == 0
     assert len(q.get_tokens(query.TokenRange(0, 1), query.TokenType.QUALIFIER)) == 1
 
     assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.PARTIAL)) == 1
-    assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.CATEGORY)) == 0
+    assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.NEAR_ITEM)) == 0
     assert len(q.get_tokens(query.TokenRange(1, 2), query.TokenType.QUALIFIER)) == 1