agent-framework-lib 0.8.1.post3__tar.gz → 0.8.1.post4__tar.gz

{agent_framework_lib-0.8.1.post3/agent_framework_lib.egg-info → agent_framework_lib-0.8.1.post4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-framework-lib
-Version: 0.8.1.post3
+Version: 0.8.1.post4
 Summary: A comprehensive Python framework for building and serving conversational AI agents with FastAPI
 Author-email: Sebastian Pavel <sebastian@cinco.ai>, Elliott Girard <elliott.girard@icloud.com>
 Maintainer-email: Sebastian Pavel <sebastian@cinco.ai>
@@ -72,6 +72,7 @@ Requires-Dist: openpyxl>=3.1.5
 Requires-Dist: python-pptx>=1.0.0
 Requires-Dist: python-docx>=1.1.0
 Requires-Dist: jsonschema>=4.20.0
+Requires-Dist: pymupdf>=1.27.2.2
 Provides-Extra: llamaindex
 Requires-Dist: llama-index>=0.14.16; extra == "llamaindex"
 Requires-Dist: llama-index-core>=0.14.16; extra == "llamaindex"
@@ -902,7 +903,7 @@ class MySkillsAgent(LlamaIndexAgent):
 |----------|--------|
 | **Visualization** | chart, mermaid, table |
 | **Document** | file, unified_pdf, file_access, excel, drawio, **powerpoint**, **word** |
-| **Web** | web_search |
+| **Web** | web_news_search |
 | **Multimodal** | multimodal, **image_gen** |
 | **UI** | form, optionsblock, image_display, **email_template**, **skill_creator** |
 | **Data** | **csv**, **data_format** |

{agent_framework_lib-0.8.1.post3/agent_framework/web/docs → agent_framework_lib-0.8.1.post4}/README.md

@@ -753,7 +753,7 @@ class MySkillsAgent(LlamaIndexAgent):
 |----------|--------|
 | **Visualization** | chart, mermaid, table |
 | **Document** | file, unified_pdf, file_access, excel, drawio, **powerpoint**, **word** |
-| **Web** | web_search |
+| **Web** | web_news_search |
 | **Multimodal** | multimodal, **image_gen** |
 | **UI** | form, optionsblock, image_display, **email_template**, **skill_creator** |
 | **Data** | **csv**, **data_format** |

{agent_framework_lib-0.8.1.post3 → agent_framework_lib-0.8.1.post4}/agent_framework/core/activity_formatter.py

@@ -161,7 +161,7 @@ SKILL_DESCRIPTIONS: dict[str, str] = {
     "pdf_with_images": "Création de PDF avec images intégrées automatiquement",
     "file": "Création, lecture et listage de fichiers",
     "file_access": "Obtention des chemins et URLs des fichiers stockés",
-    "web_search": "Recherche d'informations et d'actualités sur le web",
+    "web_news_search": "Recherche d'actualités sur le web",
     "multimodal": "Description, OCR et analyse d'images par IA",
     "form": "Génération de formulaires interactifs",
     "optionsblock": "Génération de boutons d'options interactifs",

{agent_framework_lib-0.8.1.post3 → agent_framework_lib-0.8.1.post4}/agent_framework/core/base_agent.py

@@ -578,6 +578,44 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
             logger.error(f"Error creating skill management tools: {e}")
             return []
 
+    def _get_native_skill_tools(self) -> list[callable]:
+        """Get native tools always available to the agent.
+
+        Includes:
+        - ``shell_exec`` — required by skills to run scripts.
+        - ``web_fetch`` — fetch text content from a URL.
+        - ``web_search`` — search the web via DuckDuckGo.
+
+        Returns:
+            List of callable tool functions.
+        """
+        if not SKILLS_AVAILABLE or not self._enable_skills:
+            return []
+
+        tools: list[callable] = []
+        try:
+            from ..tools.shell_tool import ShellTool
+
+            tools.append(ShellTool(default_timeout=120).get_tool_function())
+        except Exception as e:
+            logger.warning(f"Could not create ShellTool: {e}")
+
+        try:
+            from ..tools.web_fetch_tool import WebFetchTool
+
+            tools.append(WebFetchTool().get_tool_function())
+        except Exception as e:
+            logger.warning(f"Could not create WebFetchTool: {e}")
+
+        try:
+            from ..tools.web_search_tool import WebSearchTool
+
+            tools.append(WebSearchTool().get_tool_function())
+        except Exception as e:
+            logger.warning(f"Could not create WebSearchTool: {e}")
+
+        return tools
+
     async def _get_all_tools(self) -> list[callable]:
         """
         Get all tools including agent-defined tools, skill management, memory, and spawn tools.
@@ -585,8 +623,9 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
         Tool Loading Order:
         1. Agent's custom tools (from get_agent_tools())
         2. Skill management tools (list_skills, load_skill, unload_skill)
-        3. Memory tools (if memory is configured)
-        4. SubagentSpawnTool (if executor + capability_set configured)
+        3. Native skill execution tools (shell_exec, web_fetch)
+        4. Memory tools (if memory is configured)
+        5. SubagentSpawnTool (if executor + capability_set configured)
 
         Returns:
             Combined list of all available tools for this agent.
@@ -604,7 +643,15 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
                     f"🎯 Added {len(skill_mgmt_tools)} skill management tools to agent {self.agent_id}"
                 )
 
-        # 3. Add memory tools if memory is configured
+            # 3. Add native skill execution tools (shell_exec, web_fetch)
+            native_tools = self._get_native_skill_tools()
+            if native_tools:
+                tools = tools + native_tools
+                logger.info(
+                    f"🐚 Added {len(native_tools)} native skill execution tools to agent {self.agent_id}"
+                )
+
+        # 4. Add memory tools if memory is configured
         if MEMORY_AVAILABLE and hasattr(self, "_ensure_memory_initialized"):
             try:
                 # Initialize memory if configured (this is async)
@@ -619,7 +666,7 @@ class BaseAgent(SkillsMixin, MemoryMixin, AgentInterface):
             except Exception as e:
                 logger.warning(f"Failed to add memory tools: {e}")
 
-        # 4. Add subagent spawn tool if configured
+        # 5. Add subagent spawn tool if configured
         try:
             spawn_tool = self._create_subagent_spawn_tool()
             if spawn_tool is not None:

{agent_framework_lib-0.8.1.post3 → agent_framework_lib-0.8.1.post4}/agent_framework/core/step_display_config.py

@@ -634,10 +634,10 @@ DEFAULT_TOOL_DISPLAY: dict[str, StepDisplayInfo] = {
         icon="🔗",
         category="skill",
     ),
-    "skill:web_search": StepDisplayInfo(
-        id="skill:web_search",
-        friendly_name="🔍 Recherche sur Internet",
-        description="Recherche d'informations et d'actualités sur le web",
+    "skill:web_news_search": StepDisplayInfo(
+        id="skill:web_news_search",
+        friendly_name="📰 Recherche d'actualités",
+        description="Recherche d'actualités sur le web",
         icon="🔍",
         category="skill",
     ),

{agent_framework_lib-0.8.1.post3 → agent_framework_lib-0.8.1.post4}/agent_framework/skills/builtin/skills/file/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: file
-description: "Create, list, and read files from storage"
+description: "Create, list, and read files from storage (supports TXT, DOCX, XLSX, PPTX, PDF)"
 display_name: "Gestion de fichiers"
 metadata:
   config:
@@ -14,6 +14,17 @@ metadata:
       - write file
       - text file
       - store file
+      - word document
+      - excel spreadsheet
+      - powerpoint presentation
+      - docx
+      - xlsx
+      - pptx
+      - pdf
+      - read docx
+      - read xlsx
+      - read pptx
+      - read pdf
     requires:
       bins:
         - uv
@@ -23,6 +34,20 @@ metadata:
 
 Perform file operations (create, list, read) by running standalone scripts via `shell_exec`.
 
+### Supported File Formats for Reading
+
+The `read_file.py` script supports multiple file formats:
+
+| Format | Extension | Description | Extracted Content |
+|--------|-----------|-------------|-------------------|
+| **Text** | `.txt`, `.md`, `.json`, `.csv`, etc. | Plain text files | Raw text content |
+| **Word** | `.docx` | Microsoft Word documents | Paragraphs + tables |
+| **Excel** | `.xlsx` | Microsoft Excel spreadsheets | All sheets with cell values |
+| **PowerPoint** | `.pptx` | Microsoft PowerPoint presentations | Slide titles, content, notes |
+| **PDF** | `.pdf` | PDF documents | Text from all pages |
+
+**Note:** Binary files and images within documents are not extracted.
+
 ---
 
 ### 1. Create a File
@@ -144,8 +169,103 @@ uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "/p
 
 #### Output Format (stdout)
 
+The output format now includes file type and metadata:
+
+```json
+{
+  "status": "success",
+  "results": "The file content as a string...",
+  "file_type": "text",
+  "metadata": {
+    "encoding": "utf-8"
+  }
+}
+```
+
+**Metadata by file type:**
+
+| File Type | Metadata Fields |
+|-----------|----------------|
+| `text` | `encoding` |
+| `docx` | `num_paragraphs`, `num_tables` |
+| `xlsx` | `num_sheets`, `sheet_names` (list) |
+| `pptx` | `num_slides` |
+| `pdf` | `num_pages` |
+
+#### Examples by Format
+
+**Read a Word document (.docx):**
+
+```bash
+uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "report.docx"
+```
+
+Output:
 ```json
-{"status": "success", "results": "The file content as a string..."}
+{
+  "status": "success",
+  "results": "Document Title\n\nParagraph 1...\n\nParagraph 2...\n\nHeader 1 | Header 2\nValue 1 | Value 2",
+  "file_type": "docx",
+  "metadata": {
+    "num_paragraphs": 3,
+    "num_tables": 1
+  }
+}
+```
+
+**Read an Excel spreadsheet (.xlsx):**
+
+```bash
+uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "data.xlsx"
+```
+
+Output:
+```json
+{
+  "status": "success",
+  "results": "Sheet: Sheet1\nName | Age | City\nAlice | 30 | Paris\nBob | 25 | London\n\nSheet: Sheet2\nProduct | Price\nLaptop | 999.99",
+  "file_type": "xlsx",
+  "metadata": {
+    "num_sheets": 2,
+    "sheet_names": ["Sheet1", "Sheet2"]
+  }
+}
+```
+
+**Read a PowerPoint presentation (.pptx):**
+
+```bash
+uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "presentation.pptx"
+```
+
+Output:
+```json
+{
+  "status": "success",
+  "results": "Slide 1:\nIntroduction\nWelcome to our presentation\n\nSlide 2:\nAgenda\nPoint 1\nPoint 2\nNotes: Remember to explain each point",
+  "file_type": "pptx",
+  "metadata": {
+    "num_slides": 2
+  }
+}
+```
+
+**Read a PDF document (.pdf):**
+
+```bash
+uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "document.pdf"
+```
+
+Output:
+```json
+{
+  "status": "success",
+  "results": "Page 1:\nDocument Title\nContent of page 1...\n\nPage 2:\nContinued content...",
+  "file_type": "pdf",
+  "metadata": {
+    "num_pages": 2
+  }
+}
 ```
 
 #### Error Handling (stderr)
@@ -159,6 +279,7 @@ uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "/p
 | 1 | Empty file path | `"File path cannot be empty"` |
 | 1 | File not found | `"File not found: /path/to/missing.txt"` |
 | 1 | Path is a directory | `"Path is a directory, not a file: /path/to/dir"` |
+| 1 | Missing dependency | `"python-docx is required to read .docx files. Install with: uv add python-docx"` |
 | 2 | Missing required arguments | argparse usage message |
 
 ---
@@ -184,4 +305,28 @@ uv run python -m agent_framework.skills.builtin.skills.file.read_file --path "/p
 
 - All scripts output JSON to stdout on success and JSON to stderr on failure.
 - Supported file types for creation: any text-based format (.txt, .json, .md, .csv, etc.).
-- Binary files cannot be read via `read_file.py` (UTF-8 encoding is assumed).
+- Supported file types for reading: text files, DOCX, XLSX, PPTX, PDF.
+
+### Dependencies
+
+The following Python packages are **already included** in AgentFramework:
+
+| Format | Required Package | Version |
+|--------|-----------------|---------|
+| DOCX | `python-docx` | >=1.1.0 |
+| XLSX | `openpyxl` | >=3.1.5 |
+| PPTX | `python-pptx` | >=1.0.0 |
+| PDF | `pymupdf` | (optional, install if needed) |
+
+**Note:** If you get a "module not found" error, install the missing package:
+```bash
+uv add python-docx openpyxl python-pptx pymupdf
+```
+
+### Limitations
+
+- **Images**: Images within documents are not extracted (text only).
+- **Formatting**: Text formatting (bold, italic, colors) is not preserved.
+- **Formulas**: Excel formulas are evaluated and only their results are extracted.
+- **OCR**: Scanned PDFs without text layer will appear empty (OCR not supported).
+- **Password-protected files**: Encrypted documents cannot be read.

agent_framework_lib-0.8.1.post4/agent_framework/skills/builtin/skills/file/read_file.py

@@ -0,0 +1,285 @@
+#!/usr/bin/env python3
+"""Standalone CLI script for reading file contents.
+
+Reads a file from the filesystem and outputs its content
+as a JSON result on stdout.
+
+Supports multiple formats:
+- Text files (.txt, .md, .json, .csv, etc.)
+- Word documents (.docx)
+- Excel spreadsheets (.xlsx)
+- PowerPoint presentations (.pptx)
+- PDF documents (.pdf)
+"""
+import argparse
+import json
+import os
+import sys
+from pathlib import Path
+from typing import Any
+
+
+def extract_docx_text(path: str) -> tuple[str, dict[str, Any]]:
+    """Extract text from a Word document (.docx).
+
+    Args:
+        path: Path to the .docx file.
+
+    Returns:
+        Tuple of (extracted_text, metadata)
+
+    Raises:
+        ImportError: If python-docx is not installed.
+    """
+    try:
+        from docx import Document
+    except ImportError:
+        raise ImportError(
+            "python-docx is required to read .docx files. "
+            "Install with: uv add python-docx"
+        )
+
+    doc = Document(path)
+
+    # Extract paragraphs
+    paragraphs = [para.text for para in doc.paragraphs if para.text.strip()]
+
+    # Extract tables
+    table_texts = []
+    for table in doc.tables:
+        for row in table.rows:
+            row_text = " | ".join(cell.text.strip() for cell in row.cells)
+            if row_text.strip():
+                table_texts.append(row_text)
+
+    # Combine all text
+    all_text = "\n\n".join(paragraphs)
+    if table_texts:
+        all_text += "\n\n" + "\n".join(table_texts)
+
+    metadata = {
+        "num_paragraphs": len(paragraphs),
+        "num_tables": len(doc.tables),
+    }
+
+    return all_text, metadata
+
+
+def extract_xlsx_text(path: str) -> tuple[str, dict[str, Any]]:
+    """Extract text from an Excel spreadsheet (.xlsx).
+
+    Args:
+        path: Path to the .xlsx file.
+
+    Returns:
+        Tuple of (extracted_text, metadata)
+
+    Raises:
+        ImportError: If openpyxl is not installed.
+    """
+    try:
+        from openpyxl import load_workbook
+    except ImportError:
+        raise ImportError(
+            "openpyxl is required to read .xlsx files. "
+            "Install with: uv add openpyxl"
+        )
+
+    workbook = load_workbook(path, data_only=True)
+    sheets_text = []
+
+    for sheet_name in workbook.sheetnames:
+        sheet = workbook[sheet_name]
+        sheet_text = [f"Sheet: {sheet_name}"]
+
+        for row in sheet.iter_rows(values_only=True):
+            # Filter out None values and convert to strings
+            row_values = [str(cell) for cell in row if cell is not None]
+            if row_values:
+                sheet_text.append(" | ".join(row_values))
+
+        if len(sheet_text) > 1:  # More than just the sheet name
+            sheets_text.append("\n".join(sheet_text))
+
+    all_text = "\n\n".join(sheets_text)
+
+    metadata = {
+        "num_sheets": len(workbook.sheetnames),
+        "sheet_names": workbook.sheetnames,
+    }
+
+    return all_text, metadata
+
+
+def extract_pptx_text(path: str) -> tuple[str, dict[str, Any]]:
+    """Extract text from a PowerPoint presentation (.pptx).
+
+    Args:
+        path: Path to the .pptx file.
+
+    Returns:
+        Tuple of (extracted_text, metadata)
+
+    Raises:
+        ImportError: If python-pptx is not installed.
+    """
+    try:
+        from pptx import Presentation
+    except ImportError:
+        raise ImportError(
+            "python-pptx is required to read .pptx files. "
+            "Install with: uv add python-pptx"
+        )
+
+    prs = Presentation(path)
+    slides_text = []
+
+    for i, slide in enumerate(prs.slides, 1):
+        slide_text = [f"Slide {i}:"]
+
+        # Extract text from shapes
+        for shape in slide.shapes:
+            if hasattr(shape, "text") and shape.text.strip():
+                slide_text.append(shape.text.strip())
+
+        # Extract notes
+        if slide.has_notes_slide:
+            notes_text = slide.notes_slide.notes_text_frame.text.strip()
+            if notes_text:
+                slide_text.append(f"Notes: {notes_text}")
+
+        if len(slide_text) > 1:  # More than just "Slide N:"
+            slides_text.append("\n".join(slide_text))
+
+    all_text = "\n\n".join(slides_text)
+
+    metadata = {
+        "num_slides": len(prs.slides),
+    }
+
+    return all_text, metadata
+
+
+def extract_pdf_text(path: str) -> tuple[str, dict[str, Any]]:
+    """Extract text from a PDF document (.pdf).
+
+    Args:
+        path: Path to the .pdf file.
+
+    Returns:
+        Tuple of (extracted_text, metadata)
+
+    Raises:
+        ImportError: If PyMuPDF is not installed.
+    """
+    try:
+        import fitz  # PyMuPDF
+    except ImportError:
+        raise ImportError(
+            "PyMuPDF is required to read .pdf files. "
+            "Install with: uv add pymupdf"
+        )
+
+    doc = fitz.open(path)
+    pages_text = []
+
+    for page_num in range(len(doc)):
+        page = doc[page_num]
+        text = page.get_text()
+        if text.strip():
+            pages_text.append(f"Page {page_num + 1}:\n{text.strip()}")
+
+    all_text = "\n\n".join(pages_text)
+
+    metadata = {
+        "num_pages": len(doc),
+    }
+
+    doc.close()
+
+    return all_text, metadata
+
+
+def read_file(path: str) -> tuple[str, str, dict[str, Any]]:
+    """Read the contents of a file.
+
+    Args:
+        path: Path to the file to read.
+
+    Returns:
+        Tuple of (file_content, file_type, metadata)
+
+    Raises:
+        ValueError: If path is empty or file format is unsupported.
+        FileNotFoundError: If the file does not exist.
+        IsADirectoryError: If the path points to a directory.
+    """
+    if not path or not path.strip():
+        raise ValueError("File path cannot be empty")
+
+    path = os.path.abspath(path)
+
+    if not os.path.exists(path):
+        raise FileNotFoundError(f"File not found: {path}")
+
+    if os.path.isdir(path):
+        raise IsADirectoryError(f"Path is a directory, not a file: {path}")
+
+    # Detect file type
+    file_ext = Path(path).suffix.lower()
+
+    # Extract based on file type
+    try:
+        if file_ext == ".docx":
+            content, metadata = extract_docx_text(path)
+            return content, "docx", metadata
+
+        elif file_ext == ".xlsx":
+            content, metadata = extract_xlsx_text(path)
+            return content, "xlsx", metadata
+
+        elif file_ext == ".pptx":
+            content, metadata = extract_pptx_text(path)
+            return content, "pptx", metadata
+
+        elif file_ext == ".pdf":
+            content, metadata = extract_pdf_text(path)
+            return content, "pdf", metadata
+
+        else:
+            # Default: read as plain text
+            with open(path, "r", encoding="utf-8") as f:
+                content = f.read()
+            return content, "text", {"encoding": "utf-8"}
+
+    except ImportError as e:
+        raise ValueError(str(e))
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        description="Read a file and output its content as JSON on stdout."
+    )
+    parser.add_argument(
+        "--path",
+        required=True,
+        help="Path to the file to read (absolute or relative to current directory)",
+    )
+    args = parser.parse_args()
+
+    try:
+        content, file_type, metadata = read_file(args.path)
+        result = {
+            "status": "success",
+            "results": content,
+            "file_type": file_type,
+            "metadata": metadata,
+        }
+        print(json.dumps(result))
+    except Exception as e:
+        print(json.dumps({"status": "error", "message": str(e)}), file=sys.stderr)
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()