- introduce ruff as linter

- make ruff and mypy green

Signed-off-by: Florian Schepers <[email protected]>

meminto/chunking.py

@@ -38,7 +38,7 @@ def _number_of_tokens_per_chunk(
         number_of_chunks = token_count_transcript // token_count_per_chunk + 1
         number_of_tokens_per_chunk = token_count_transcript // number_of_chunks + 1
 
-        print(f"Spliting transcript in chunks:")
+        print("Spliting transcript in chunks:")
         print(f"LLM max. token count: {max_tokens}")
         print(f"Token count of system prompt: {token_count_system_prompt}")
         print(f"Token count reserved for response: {token_count_reserved_for_response}")

meminto/diarizer.py

@@ -1,6 +1,6 @@
 from pathlib import Path
-from pyannote.audio import Pipeline # type: ignore
-from pyannote.core import Annotation # type: ignore
+from pyannote.audio import Pipeline 
+from pyannote.core import Annotation 
 from meminto.decorators import log_time
 
 

meminto/helpers.py

@@ -32,7 +32,7 @@ def parse_input_file_path(input_file: str) -> Path:
     if not file_path.is_file():
         raise Exception(f"Input file path '{file_path}' does not reference a file.")
 
-    if not file_path.suffix in ALLOWED_INPUT_FILE_TYPE:
+    if file_path.suffix not in ALLOWED_INPUT_FILE_TYPE:
         raise Exception(
             f"Invalid input file type. Only one of the following file type are allowed: {', '.join(str(file_type) for file_type in ALLOWED_INPUT_FILE_TYPE)}"
         )

meminto/llm/llm.py

@@ -1,4 +1,4 @@
-import requests  # type: ignore
+import requests
 
 
 class LLM:

meminto/llm/tokenizers.py

@@ -1,6 +1,6 @@
 import tiktoken
 from transformers import AutoTokenizer, OpenAIGPTTokenizer
-from huggingface_hub import login
+from huggingface_hub import login 
 
 
 class Tokenizer:
@@ -13,7 +13,7 @@ def _select_tokenizer(self):
         login(token=self.hugging_face_acces_token)
         try:
             tokenizer = AutoTokenizer.from_pretrained(self.model)
-        except:
+        except(Exception):
             if self.model in tiktoken.model.MODEL_TO_ENCODING.keys():
                 tokenizer = OpenAIGPTTokenizer.from_pretrained("openai-gpt")
             else:

meminto/meeting_minutes_generator.py

@@ -1,5 +1,3 @@
-import os
-from typing import Tuple
 from meminto.decorators import log_time
 from meminto.helpers import Language
 from meminto.llm.llm import LLM
@@ -16,7 +14,7 @@
 )
 from meminto.llm.tokenizers import Tokenizer
 from meminto.transcriber import TranscriptSection
-from huggingface_hub import login
+
 
 
 class MeetingMinutesGenerator:

pyproject.toml

@@ -21,7 +21,19 @@ transformers = "^4.39.3"
 pyannote-pipeline = "^3.0.1"
 speechbrain = "0.5.15"
 python-dotenv = "^1.0.1"
+
+[tool.poetry.group.dev.dependencies]
+ruff = "^0.4.2"
 mypy = "^1.9.0"
+types-requests = "^2.31.0.20240406"
 
+[[tool.mypy.overrides]]
+module = [
+    "transformers",
+    "huggingface_hub",
+    "torchaudio",
+    "pyannote.*",
+]
+ignore_missing_imports = true