From 655d4a1a8dba4fc810ce4ab725e73b3dee3b70cf Mon Sep 17 00:00:00 2001
From: Tobias Wochinger <tobias.wochinger@deepset.ai>
Date: Tue, 5 Mar 2024 12:14:10 +0100
Subject: [PATCH] test: test for missing dependencies (#7278)

* tests: import test for missing libraries

* build: add missing dependencies

* refactor: use glob instead of tree walk

* test: extract constants + more documentation
---
 haystack/testing/document_store.py |  5 ++-
 pyproject.toml                     |  7 ++++
 test/test_imports.py               | 64 ++++++++++++++++++++++++++++++
 3 files changed, 75 insertions(+), 1 deletion(-)
 create mode 100644 test/test_imports.py

diff --git a/haystack/testing/document_store.py b/haystack/testing/document_store.py
index 77b1659015..0cdfdae042 100644
--- a/haystack/testing/document_store.py
+++ b/haystack/testing/document_store.py
@@ -4,12 +4,15 @@
 from typing import List
 
 import pandas as pd
-import pytest
 
 from haystack.dataclasses import Document
 from haystack.document_stores.errors import DuplicateDocumentError
 from haystack.document_stores.types import DocumentStore, DuplicatePolicy
 from haystack.errors import FilterError
+from haystack.lazy_imports import LazyImport
+
+with LazyImport("Run 'pip install pytest'") as pytest_import:
+    import pytest
 
 
 def _random_embeddings(n):
diff --git a/pyproject.toml b/pyproject.toml
index 22b23d009f..393b9caa28 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -59,6 +59,9 @@ dependencies = [
   "networkx", # Pipeline graphs
   "typing_extensions>=4.7", # typing support for Python 3.8
   "boilerpy3", # Fulltext extraction from HTML pages
+  "requests",
+  "numpy",
+  "python-dateutil",
 ]
 
 [tool.hatch.envs.default]
@@ -127,6 +130,10 @@ extra-dependencies = [
 
   # Structured logging
   "structlog",
+
+  # Looking for missing imports
+  "isort",
+  "pyproject-parser",
 ]
 
 [tool.hatch.envs.test.scripts]
diff --git a/test/test_imports.py b/test/test_imports.py
new file mode 100644
index 0000000000..5cf59f23ff
--- /dev/null
+++ b/test/test_imports.py
@@ -0,0 +1,64 @@
+import ast
+import os
+from _ast import Import, ImportFrom
+from pathlib import Path
+
+import isort
+import toml
+from pyproject_parser import PyProject
+
+# Some libraries have different names in the import and in the dependency
+# If below test fails due to that, add the library name to the dictionary
+LIBRARY_NAMES_TO_MODULE_NAMES = {"python-dateutil": "dateutil"}
+
+# Some standard libraries are not detected by isort. If below test fails due to that, add the library name to the set.
+ADDITIONAL_STD_LIBS = {"yaml"}
+
+
+def test_for_missing_dependencies() -> None:
+    # We implement this manual check because
+    # - All tools out there are too powerful because they find all the imports in the haystack package
+    # - if we import all modules to check the imports we don't find issues with direct dependencies which are also
+    #   sub-dependencies of other dependencies
+
+    #### Collect imports
+    top_level_imports = set()
+    for path in Path("haystack").glob("**/*.py"):
+        content = path.read_text(encoding="utf-8")
+        tree = ast.parse(content)
+        for item in tree.body:
+            if isinstance(item, Import):
+                module = item.names[0].name
+            elif isinstance(item, ImportFrom) and item.level == 0:  # level > 1 are relative imports
+                module = item.module
+            else:
+                # we only care about imports
+                break
+
+            top_level_imports.add(module.split(".")[0])
+
+    third_party_modules = {
+        module
+        for module in top_level_imports
+        if isort.place_module(module) == "THIRDPARTY" and module not in ADDITIONAL_STD_LIBS
+    }
+
+    #### Load specified dependencies
+    parsed = toml.load("pyproject.toml")
+    # Pyproject complains about our pyproject.toml file, so we need to parse only the dependencies
+    # We still need `PyProject` to parse the dependencies (think of markers and stuff)
+    only_dependencies = {"project": {"name": "test", "dependencies": parsed["project"]["dependencies"]}}
+    project_dependencies = PyProject.project_table_parser.parse(only_dependencies["project"], set_defaults=True)[
+        "dependencies"
+    ]
+
+    project_dependency_modules = set()
+    for dep in project_dependencies:
+        if dep.name in LIBRARY_NAMES_TO_MODULE_NAMES:
+            project_dependency_modules.add(LIBRARY_NAMES_TO_MODULE_NAMES[dep.name])
+
+        project_dependency_modules.add(dep.name.replace("-", "_"))
+
+    #### And now finally; the check
+    for module in third_party_modules:
+        assert module in project_dependency_modules, f"Module {module} is not in the dependencies"