first transition to clean tools

.gitignore

@@ -1,3 +1,2 @@
 .env
-pg.py
 *.json

agent.py

@@ -1,6 +1,11 @@
 import os
-from smolagents import CodeAgent, tool, DuckDuckGoSearchTool, OpenAIServerModel, VisitWebpageTool, Tool
+from smolagents import CodeAgent, tool, DuckDuckGoSearchTool, OpenAIServerModel, VisitWebpageTool, PythonInterpreterTool
 from langchain_community.document_loaders import WikipediaLoader, ArxivLoader
+from tools.csv_reader import CsvReaderTool
+from tools.excel_reader import ExcelReaderTool
+from tools.file_downloader import FileDownloaderTool
+from tools.img_txt_extractor import ImageTextExtractorTool
+from tools.pdf_reader import PdfReaderTool
 
 @tool
 def add(a:int, b:int) -> int:
@@ -113,6 +118,13 @@ def arvix_search(query: str) -> str:
 def get_agent() -> CodeAgent:
   search_tool = DuckDuckGoSearchTool()
   web_page_tool = VisitWebpageTool()
+  py_interpreter = PythonInterpreterTool()
+
+  csv_reader = CsvReaderTool()
+  excel_reader = ExcelReaderTool()
+  file_downloader = FileDownloaderTool()
+  img_txt_extractor = ImageTextExtractorTool()
+  pdf_reader = PdfReaderTool()
 
   api_key = os.getenv('CODESTRAL_API_KEY')
 
@@ -121,4 +133,4 @@ def get_agent() -> CodeAgent:
     api_base="https://codestral.mistral.ai/v1/",
     api_key=api_key)
 
-  return CodeAgent(tools=[add, subtract, multiply, divide, modulus, rounder, search_tool, web_page_tool, wiki_search, arvix_search], model=model)
+  return CodeAgent(tools=[add, subtract, multiply, divide, modulus, rounder, search_tool, web_page_tool, py_interpreter, wiki_search, arvix_search, csv_reader, excel_reader, file_downloader, img_txt_extractor, pdf_reader], model=model)

requirements.txt

@@ -1,10 +1,11 @@
 gradio
 requests
+pandas
+smolagents
 smolagents[openai]
-langchain_community
-wikipedia
 duckduckgo_search
 requests
 markdownify
-arxiv
-pymupdf
+openpyxl
+easyocr
+pdfminer.six

tools/csv_reader.py

@@ -0,0 +1,21 @@
+import pandas as pd
+from smolagents import Tool
+
+class CsvReaderTool(Tool):
+    name = "csv_reader"
+    description = "Extract CSV file content. Supported file extensions: .csv"
+    inputs = {
+        "file_path": {
+            "type": "string",
+            "description": "Path to the CSV file",
+        }
+    }
+    output_type = "string"
+
+    def forward(self, file_path) -> str:
+        try:
+            df = pd.read_csv(file_path)
+            print(f"Describe CSV file:\n {df.describe()}")
+            return df.to_json()
+        except Exception as e:
+            return f"Error processing CSV file: {str(e)}"

tools/excel_reader.py

@@ -0,0 +1,21 @@
+import pandas as pd
+from smolagents import Tool
+
+class ExcelReaderTool(Tool):
+  name = "excel_reader"
+  description = "Extract Excel file content. Supported file extensions: .xls, .xlsx, .xlsb, .xlsm, .odf, .ods, .odt"
+  inputs = {
+        "file_path": {
+            "type": "string",
+            "description": "Path to the Excel file",
+        }
+    }
+  output_type = "string"
+
+  def forward(self, file_path) -> str:
+      try:
+          df = pd.read_excel(file_path)
+          print(f"Describe Excel file:\n {df.describe()}")
+          return df.to_json()
+      except Exception as e:
+          return f"Error processing Excel file: {str(e)}"

tools/file_downloader.py

@@ -0,0 +1,24 @@
+import urllib.request
+from smolagents import Tool
+
+class FileDownloaderTool(Tool):
+    name = "file_downloader"
+    description = "Download a file from Internet by URL provided, save it into temp dir and return file path"
+    inputs = {
+        "url": {
+            "type": "string",
+            "description": "URL to download from",
+        }
+    }
+    output_type = "string"
+
+    def forward(self, url: str) -> str:
+        file_path = None
+
+        try:
+            result = urllib.request.urlretrieve(url)
+            file_path = result[0]
+        except Exception as e:
+            print(f"Error downloading file: {str(e)}")
+
+        return file_path

tools/img_txt_extractor.py

@@ -0,0 +1,24 @@
+from smolagents import Tool
+import easyocr
+
+class ImageTextExtractorTool(Tool):
+    name = "img_txt_extractor"
+    description = """
+        Multilingual OCR tool to extract key information or presented text from any image.
+        Supported image extensions: .png, .jpg, .jpeg, .bmp, .svg.
+    """
+    inputs = {
+        "image_path": {
+            "type": "string",
+            "description": "The path to the image file",
+        }
+    }
+    output_type = "array"
+
+    def __init__(self):
+        super().__init__()
+        self.reader = easyocr.Reader(['ch_sim', 'en'])
+
+    def forward(self, image_path: str) -> list[str]:
+        result = self.reader.readtext(image_path, detail=False)
+        return result

tools/pdf_reader.py

@@ -0,0 +1,19 @@
+from pdfminer.high_level import extract_text
+from smolagents import Tool
+
+class PdfReaderTool(Tool):
+    name = "pdf_reader"
+    description = "Extract PDF content. Supported file extensions: .pdf"
+    inputs = {
+        "file_path": {
+            "type": "string",
+            "description": "Path to the PDF file",
+        }
+    }
+    output_type = "string"
+
+    def forward(self, file_path) -> str:
+        try:
+            return extract_text(file_path)
+        except Exception as e:
+            return f"Error processing PDF file: {str(e)}"