Update app.py
Browse files
app.py
CHANGED
|
@@ -3,7 +3,7 @@ import gradio as gr
|
|
| 3 |
import requests
|
| 4 |
import inspect
|
| 5 |
import pandas as pd
|
| 6 |
-
from smolagents import CodeAgent, InferenceClientModel, DuckDuckGoSearchTool,
|
| 7 |
|
| 8 |
|
| 9 |
# (Keep Constants as is)
|
|
@@ -75,19 +75,22 @@ class BasicAgent:
|
|
| 75 |
"""
|
| 76 |
if file_path.endswith(".csv"):
|
| 77 |
df = pd.read_csv(file_path)
|
| 78 |
-
elif file_path.edswith(".pdf"):
|
| 79 |
-
df = pd.read_pdf(file_path)
|
| 80 |
else:
|
| 81 |
df = pd.read_excel(file_path)
|
| 82 |
return df.to_string()
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 91 |
#pdf_tool = load_tool(
|
| 92 |
#"matterattetatte/pdf-upload-extractor-tool",
|
| 93 |
# trust_remote_code = True
|
|
@@ -100,10 +103,12 @@ class BasicAgent:
|
|
| 100 |
FinalAnswerTool(),
|
| 101 |
#pdf_tool,
|
| 102 |
fetch_task_file,
|
|
|
|
| 103 |
read_spreadsheet,
|
| 104 |
get_youtube_transcript,
|
| 105 |
],
|
| 106 |
model= InferenceClientModel("Qwen/Qwen2.5-Coder-32B-Instruct"),
|
|
|
|
| 107 |
#instructions = "If the question includes a task_id and mentions a file, call fetch_task_file first; route YouTube URLs to get_youtube_transcript, other URLs to visit_webpage, PDFs to pdf_tool, and spreadsheets to read_spreadsheet, using web_search only when no URL or file is given,then respond with only the exact final answer value, no explanation, no prefix.",
|
| 108 |
max_steps=20,
|
| 109 |
)
|
|
|
|
| 3 |
import requests
|
| 4 |
import inspect
|
| 5 |
import pandas as pd
|
| 6 |
+
from smolagents import CodeAgent, InferenceClientModel, DuckDuckGoSearchTool,Tool,tool,PythonInterpreterTool,VisitWebpageTool,FinalAnswerTool,load_tool
|
| 7 |
|
| 8 |
|
| 9 |
# (Keep Constants as is)
|
|
|
|
| 75 |
"""
|
| 76 |
if file_path.endswith(".csv"):
|
| 77 |
df = pd.read_csv(file_path)
|
|
|
|
|
|
|
| 78 |
else:
|
| 79 |
df = pd.read_excel(file_path)
|
| 80 |
return df.to_string()
|
| 81 |
+
|
| 82 |
+
@tool
|
| 83 |
+
def read_pdf(file_path: str) -> str:
|
| 84 |
+
"""
|
| 85 |
+
Extracts text from a PDF file.
|
| 86 |
+
Args:
|
| 87 |
+
file_path: Local path to the PDF file.
|
| 88 |
+
Returns:
|
| 89 |
+
The extracted text content
|
| 90 |
+
"""
|
| 91 |
+
from pypdf import PdfReader
|
| 92 |
+
reader = PdfReader(file_path)
|
| 93 |
+
return "\n".join(page.extract_text() or "" for page in reader.pages)
|
| 94 |
#pdf_tool = load_tool(
|
| 95 |
#"matterattetatte/pdf-upload-extractor-tool",
|
| 96 |
# trust_remote_code = True
|
|
|
|
| 103 |
FinalAnswerTool(),
|
| 104 |
#pdf_tool,
|
| 105 |
fetch_task_file,
|
| 106 |
+
read_pdf,
|
| 107 |
read_spreadsheet,
|
| 108 |
get_youtube_transcript,
|
| 109 |
],
|
| 110 |
model= InferenceClientModel("Qwen/Qwen2.5-Coder-32B-Instruct"),
|
| 111 |
+
additional_authorized_imports=["pandas", "requests", "re"],
|
| 112 |
#instructions = "If the question includes a task_id and mentions a file, call fetch_task_file first; route YouTube URLs to get_youtube_transcript, other URLs to visit_webpage, PDFs to pdf_tool, and spreadsheets to read_spreadsheet, using web_search only when no URL or file is given,then respond with only the exact final answer value, no explanation, no prefix.",
|
| 113 |
max_steps=20,
|
| 114 |
)
|