Upload 4 files

tools.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 import contextlib
 import io
 import os
-from typing import Any, Dict, List, Hashable
+from typing import Any, Dict, List
 
 from smolagents import Tool
 
@@ -14,8 +14,11 @@ class PythonRunTool(Tool):
         "Execute trusted Python code and return printed output "
         "+ repr() of the last expression (or _result variable)."
     )
+    inputs = {
+        "code": {"type": str, "description": "Python code to execute", "required": True}
+    }
 
-    def forward(self, code: str) -> str:  # type: ignore[override]
+    def forward(self, code: str) -> str:
         buf, ns = io.StringIO(), {}
         last = None
         try:
@@ -34,8 +37,17 @@ class ExcelLoaderTool(Tool):
         "Read .xlsx/.xls/.csv from disk and return "
         "rows as a list of dictionaries with string keys."
     )
+    inputs = {
+        "path": {"type": str, "description": "Path to .csv/.xls/.xlsx file", "required": True},
+        "sheet": {
+            "type": str,
+            "description": "Sheet name or index (optional, required for Excel files only)",
+            "required": False,
+            "default": None,
+        }
+    }
 
-    def forward(self, path: str, sheet: str | int | None = None) -> List[Dict[str, Any]]:  # type: ignore[override]
+    def forward(self, path: str, sheet: str | int | None = None) -> List[Dict[str, Any]]:
         import pandas as pd
         if not os.path.isfile(path):
             raise FileNotFoundError(path)
@@ -44,7 +56,6 @@ class ExcelLoaderTool(Tool):
             df = pd.read_csv(path)
         else:
             df = pd.read_excel(path, sheet_name=sheet)
-        # Ensure all keys are str for type safety
         records = [{str(k): v for k, v in row.items()} for row in df.to_dict(orient="records")]
         return records
 
@@ -52,10 +63,13 @@ class ExcelLoaderTool(Tool):
 class YouTubeTranscriptTool(Tool):
     name = "youtube_transcript"
     description = "Return the subtitles of a YouTube URL using youtube-transcript-api."
+    inputs = {
+        "url": {"type": str, "description": "YouTube URL", "required": True},
+        "lang": {"type": str, "description": "Transcript language (default: en)", "required": False, "default": "en"}
+    }
 
-    def forward(self, url: str, lang: str = "en") -> str:  # type: ignore[override]
+    def forward(self, url: str, lang: str = "en") -> str:
         from urllib.parse import urlparse, parse_qs
-        # Per Pylance, import from private API
         from youtube_transcript_api._api import YouTubeTranscriptApi
         vid = parse_qs(urlparse(url).query).get("v", [None])[0] or url.split("/")[-1]
         data = YouTubeTranscriptApi.get_transcript(vid, languages=[lang, "en", "en-US", "en-GB"])
@@ -65,29 +79,33 @@ class YouTubeTranscriptTool(Tool):
 class AudioTranscriptionTool(Tool):
     name = "transcribe_audio"
     description = "Transcribe an audio file with OpenAI Whisper, returns plain text."
+    inputs = {
+        "path": {"type": str, "description": "Path to audio file", "required": True},
+        "model": {"type": str, "description": "Model name for transcription (default: whisper-1)", "required": False, "default": "whisper-1"}
+    }
 
-    def forward(self, path: str, model: str = "whisper-1") -> str:  # type: ignore[override]
+    def forward(self, path: str, model: str = "whisper-1") -> str:
         import openai
-        import os
         if not os.path.isfile(path):
             raise FileNotFoundError(path)
         openai.api_key = os.getenv("OPENAI_API_KEY")
-        # Version/API guard for openai.Audio
         if not hasattr(openai, "Audio"):
             raise ImportError(
                 "Your OpenAI package does not support Audio. "
                 "Please upgrade it with: pip install --upgrade openai"
             )
         with open(path, "rb") as fp:
-            # type: ignore[attr-defined]
             return openai.Audio.transcribe(model=model, file=fp)["text"].strip()
 
 # ---- 5. SimpleOCRTool ------------------------------------------------------
 class SimpleOCRTool(Tool):
     name = "image_ocr"
     description = "Return any text spotted in an image via pytesseract OCR."
+    inputs = {
+        "path": {"type": str, "description": "Path to image file", "required": True}
+    }
 
-    def forward(self, path: str) -> str:  # type: ignore[override]
+    def forward(self, path: str) -> str:
         from PIL import Image
         import pytesseract
         if not os.path.isfile(path):
@@ -101,4 +119,4 @@ __all__ = [
     "YouTubeTranscriptTool",
     "AudioTranscriptionTool",
     "SimpleOCRTool",
-]
+]