Merge branch 'main' into image

sahajsoft · Jun 4, 2024 · 57fce6d · 57fce6d
2 parents bb7e44d + f1ece7d
commit 57fce6d
Show file tree

Hide file tree

Showing 15 changed files with 358 additions and 279 deletions.
diff --git a/.github/workflows/test.yaml b/.github/workflows/test.yaml
@@ -32,5 +32,5 @@ jobs:
       shell: bash
 
     - name: "Test"
-      run: 'shopt -s globstar && poetry run python -m unittest tests/**/*.py'
+      run: 'poetry run pytest'
       shell: bash
diff --git a/flake.nix b/flake.nix
@@ -11,7 +11,7 @@
   outputs = { self, nixpkgs, flake-utils, poetry2nix }:
     flake-utils.lib.eachDefaultSystem (system:
       let
-        nativeBuildInputs = with pkgs; [ stdenv python3 poetry ];
+        nativeBuildInputs = with pkgs; [ stdenv python3 poetry tesseract ];
         buildInputs = with pkgs; [ ];
 
         # see https://github.com/nix-community/poetry2nix/tree/master#api for more functions and examples.

diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -13,10 +13,20 @@ spacy = "^3.7.4"
 scipy = "<1.13.0"
 presidio-anonymizer = "^2.2.354"
 presidio-analyzer = {version = "^2.2.354", extras = ["transformers", "stanza"]}
+
 presidio_image_redactor = {version="^0.0.52"}
 pytesseract = {version="^0.3.10"}
 
+pytest = "^8.2.1"
+
+
 
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
+
+[tool.pytest.ini_options]
+pythonpath = "src"
+addopts = [
+    "--import-mode=importlib",
+]
diff --git a/analyzer_engine/__init__.py → src/analyzer_engine/__init__.py b/analyzer_engine/__init__.py → src/analyzer_engine/__init__.py
diff --git a/analyzer_engine/csv_analyzer_engine.py → src/analyzer_engine/csv_analyzer_engine.py b/analyzer_engine/csv_analyzer_engine.py → src/analyzer_engine/csv_analyzer_engine.py
diff --git a/config/__init__.py → src/config/__init__.py b/config/__init__.py → src/config/__init__.py
diff --git a/config/nlp_engine_config.py → src/config/nlp_engine_config.py b/config/nlp_engine_config.py → src/config/nlp_engine_config.py
diff --git a/recognizer/__init__.py → src/recognizer/__init__.py b/recognizer/__init__.py → src/recognizer/__init__.py
diff --git a/recognizer/flair_recognizer.py → src/recognizer/flair_recognizer.py b/recognizer/flair_recognizer.py → src/recognizer/flair_recognizer.py
diff --git a/tests/__init__.py b/tests/__init__.py
diff --git a/tests/analyzer_engine/__init__.py b/tests/analyzer_engine/__init__.py
diff --git a/tests/analyzer_engine/csv_analyzer_engine_test.py b/tests/analyzer_engine/csv_analyzer_engine_test.py
@@ -1,19 +1,14 @@
-import unittest
+import pytest
 
 from analyzer_engine.csv_analyzer_engine import CSVAnalyzerEngine
 from config.nlp_engine_config import FlairNLPEngine
 
 
-class CSVAnalayzerEngineTest(unittest.TestCase):
-    def setUp(self) -> None:
-        nlp_engine = FlairNLPEngine("flair/ner-english-large")
-        self.csv_analyser = CSVAnalyzerEngine(nlp_engine)
-
-    def test_csv_analyzer_engine_anonymizer(self):
-
-        from presidio_anonymizer import BatchAnonymizerEngine
-        analyzer_results = self.csv_analyser.analyze_csv('./data/sample_data.csv', language="en")
-
-        anonymizer = BatchAnonymizerEngine()
-        anonymized_results = anonymizer.anonymize_dict(analyzer_results)
-        self.assertIsNotNone(anonymized_results)
+def test_csv_analyzer_engine_anonymizer():
+    nlp_engine = FlairNLPEngine("flair/ner-english-large")
+    csv_analyzer = CSVAnalyzerEngine(nlp_engine)
+    from presidio_anonymizer import BatchAnonymizerEngine
+    analyzer_results = csv_analyzer.analyze_csv('./data/sample_data.csv', language="en")
+    anonymizer = BatchAnonymizerEngine()
+    anonymized_results = anonymizer.anonymize_dict(analyzer_results)
+    assert anonymized_results
diff --git a/tests/recognizer/__init__.py b/tests/recognizer/__init__.py
diff --git a/tests/recognizer/flair_recognizer_test.py b/tests/recognizer/flair_recognizer_test.py
@@ -1,20 +1,17 @@
-import unittest
+import pytest
 
 from recognizer.flair_recognizer import FlairRecognizer
 
 
-class TestFlairRecognizer(unittest.TestCase):
-    def setUp(self) -> None:
-        self.recognizer = FlairRecognizer(model_path="flair/ner-english-large")
-
-    def test_flair_recognizer_analyse(self):
-        test_data = "Sowmya is working in Berkley bank as an accountant since 2021"
-        result = self.recognizer.analyze(test_data)
-        self.assertGreater(len(result), 0)
-
-    def test_flair_recognizes_persons_correctly(self):
-        test_data = "Sowmya is a person name"
-        self.assertGreater(len(self.recognizer.analyze(test_data)), 0)
-        test_data = "XXXXXX is a valid name?"
-        self.assertEquals(len(self.recognizer.analyze(test_data)), 0)
+def test_flair_recognizer_analyze():
+    recognizer = FlairRecognizer(model_path="flair/ner-english-large")
+    test_data = "Sowmya is working in Berkley bank as an accountant since 2021"
+    result = recognizer.analyze(test_data)
+    assert len(result) > 0
 
+def test_flair_recognizes_persons_correctly():
+    recognizer = FlairRecognizer(model_path="flair/ner-english-large")
+    test_data = "Sowmya is a person name"
+    assert len(recognizer.analyze(test_data)) > 0
+    test_data = "XXXXXX is a valid name?"
+    assert len(recognizer.analyze(test_data)) == 0