meshagent-markitdown 0.0.37__py3-none-any.whl → 0.0.39__py3-none-any.whl

meshagent/markitdown/__init__.py

@@ -1 +1,3 @@
-from .version import __version__
+from .version import __version__
+
+__all__ = [__version__]

meshagent/markitdown/tools/__init__.py

@@ -1 +1,3 @@
-from .markitdown import MarkItDownToolkit
+from .markitdown import MarkItDownToolkit
+
+__all__ = [MarkItDownToolkit]

meshagent/markitdown/tools/markitdown.py

@@ -1,18 +1,17 @@
-import fal_client
 import aiohttp
 import mimetypes
-import uuid
-import base64
 from typing import Optional
-from urllib.parse import urlparse
-from openapi_core import OpenAPI
-import json
 import os
-from meshagent.api import EmptyResponse, JsonResponse, FileResponse
-from meshagent.tools import Tool, Toolkit, ToolContext, TextResponse, get_bytes_from_url, BlobStorage, RemoteToolkit
-from copy import deepcopy
+from meshagent.api import EmptyResponse, FileResponse
+from meshagent.tools import (
+    Tool,
+    ToolContext,
+    TextResponse,
+    get_bytes_from_url,
+    BlobStorage,
+    RemoteToolkit,
+)
 import logging
-import urllib.parse
 import asyncio
 import aiofiles
 import markitdown
@@ -20,44 +19,41 @@ import markitdown
 logger = logging.getLogger("markitdown")
 
 
-
-supported_extensions = { 
-  ".pdf",
-  ".docx",
-  ".pptx",
-  ".docx",
-  ".heic",
-  ".xlsx",
-  # TODO: actually supports more formats, do we want others?  
+supported_extensions = {
+    ".pdf",
+    ".docx",
+    ".pptx",
+    ".docx",
+    ".heic",
+    ".xlsx",
+    # TODO: actually supports more formats, do we want others?
 }
 
+
 class FileMarkItDownTool(Tool):
     def __init__(self):
         super().__init__(
-            name = "markitdown_from_file",
-            title = "MarkItDown File Adapter",
-            description="Read the contents of a PDF or Office document from a file path",   
-            input_schema = {
-                "type" : "object",
-                "additionalProperties" : False,
-                "required" : [ "path" ],
-                "properties" : {
-                    "path" : {
-                        "type" : "string"
-                    }
-                }
-            })
-        
-    async def execute(self, *, context: ToolContext, path: str):
+            name="markitdown_from_file",
+            title="MarkItDown File Adapter",
+            description="Read the contents of a PDF or Office document from a file path",
+            input_schema={
+                "type": "object",
+                "additionalProperties": False,
+                "required": ["path"],
+                "properties": {"path": {"type": "string"}},
+            },
+        )
 
+    async def execute(self, *, context: ToolContext, path: str):
         filename, ext = os.path.splitext(path)
         if ext in supported_extensions:
-            file : FileResponse = await context.room.storage.download(path=path)
+            file: FileResponse = await context.room.storage.download(path=path)
             logger.info("adding office metadata for file: {path}".format(path=path))
-            async with aiofiles.tempfile.NamedTemporaryFile('wb', suffix=ext) as f:
+            async with aiofiles.tempfile.NamedTemporaryFile("wb", suffix=ext) as f:
                 await f.write(file.data)
                 logger.info("tmp: {path}".format(path=f.name))
                 converter = markitdown.MarkItDown()
+
                 def convert():
                     return converter.convert(f.name)
 
@@ -69,120 +65,55 @@ class FileMarkItDownTool(Tool):
 
 
 class UrlMarkItDownTool(Tool):
-    def __init__(self, blob_storage: Optional[BlobStorage]  = None):
+    def __init__(self, blob_storage: Optional[BlobStorage] = None):
         super().__init__(
-            name = "markitdown_from_url",
-            title = "MarkItDown URL Adapter",
-            description = "Read the contents of a PDF or Office document from a URL",
-            input_schema = {
-                "type" : "object",
-                "additionalProperties" : False,
-                "required" : [ "url" ],
-                "properties" : {
-                    "url" : {
-                        "type" : "string"
-                    }
-                }
-            })
+            name="markitdown_from_url",
+            title="MarkItDown URL Adapter",
+            description="Read the contents of a PDF or Office document from a URL",
+            input_schema={
+                "type": "object",
+                "additionalProperties": False,
+                "required": ["url"],
+                "properties": {"url": {"type": "string"}},
+            },
+        )
 
         self._blob_storage = blob_storage
         self._session = aiohttp.ClientSession()
-        
+
     async def execute(self, *, context: ToolContext, url: str):
-        
         blob = await get_bytes_from_url(url=url, blob_storage=self._blob_storage)
-        
+
         ext = mimetypes.guess_extension(blob.mime_type)
         if ext in supported_extensions:
-            async with aiofiles.tempfile.NamedTemporaryFile('wb', suffix=ext) as f:
-
+            async with aiofiles.tempfile.NamedTemporaryFile("wb", suffix=ext) as f:
                 # TODO: should protect against too large files with maximum file length?
                 await f.write(blob.data)
-                
+
                 converter = markitdown.MarkItDown()
+
                 def convert():
                     return converter.convert(f.name)
 
-                result = await  asyncio.get_event_loop().run_in_executor(None, convert)
+                result = await asyncio.get_event_loop().run_in_executor(None, convert)
 
                 return TextResponse(text=result.text_content)
         else:
-            raise Exception("Unsupported file type, you cannot use this tool to retreive its content")
-        
+            raise Exception(
+                "Unsupported file type, you cannot use this tool to retreive its content"
+            )
 
-class AskUserMarkItDownTool(Tool):
-    def __init__(self):
-        super().__init__(
-            name = "markitdown_from_user",
-            title = "Read a file from a user",
-            description = "Read the contents of a PDF or Office document the user. Requires ask_user_file tool to be available at runtime",
-            input_schema = {
-                "type": "object",
-                "additionalProperties": False,
-                "required": ["title", "description"],
-                "properties": {
-                    "title": {
-                        "type": "string",
-                        "description": "a very short description suitable for a dialog title",
-                    },
-                    "description": {
-                        "type": "string",
-                        "description":
-                            "helpful information that explains why this information is being collected and how it will be used",
-                    },
-                },
-            }
-        )
-        
-    async def execute(self, *, context: ToolContext, title: str, description: str):
-        
-        who = context.caller
-        if context.on_behalf_of != None:
-            who = context.on_behalf_of
-
-        file_response : FileResponse = await context.room.agents.invoke_tool(
-            participant_id=who.id,
-            toolkit="ui",
-            tool="ask_user_for_file",
-            arguments={
-                "title": title,
-                "description":description
-            }
-        )
-
-        
-        ext = mimetypes.guess_extension(file_response.mime_type)
-
-        logger.info(f"got file: {file_response.mime_type} {ext}")
-        
-        if ext in supported_extensions:
-            async with aiofiles.tempfile.NamedTemporaryFile('wb', suffix=ext) as f:
-
-                # TODO: should protect against too large files with maximum file length?
-                await f.write(file_response.data)
-                
-                converter = markitdown.MarkItDown()
-                def convert():
-                    return converter.convert(f.name)
-
-                result = await  asyncio.get_event_loop().run_in_executor(None, convert)
-
-                return JsonResponse(json={
-                    "filename" : file_response.name,
-                    "mime_type" : file_response.mime_type,
-                    "content" : result.text_content
-                })
-        else:
-            raise Exception("Unsupported file type, you cannot use this tool to retreive its content")
 
 class MarkItDownToolkit(RemoteToolkit):
-    def __init__(self, blob_storage: Optional[BlobStorage] = None, name = "meshagent.markitdown"):
+    def __init__(
+        self, blob_storage: Optional[BlobStorage] = None, name="meshagent.markitdown"
+    ):
         super().__init__(
             name=name,
             title="markitdown",
-            description="MarkItDown is a utility for converting various files to Markdown",   
+            description="MarkItDown is a utility for converting various files to Markdown",
             tools=[
                 FileMarkItDownTool(),
                 UrlMarkItDownTool(blob_storage=blob_storage),
-                AskUserMarkItDownTool(),
-            ])
+            ],
+        )

meshagent/markitdown/version.py

@@ -1 +1 @@
-__version__ = "0.0.37"
+__version__ = "0.0.39"

{meshagent_markitdown-0.0.37.dist-info → meshagent_markitdown-0.0.39.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: meshagent-markitdown
-Version: 0.0.37
+Version: 0.0.39
 Summary: Markitdown support for Meshagent
 License-Expression: Apache-2.0
 Project-URL: Documentation, https://docs.meshagent.com
@@ -9,9 +9,9 @@ Project-URL: Source, https://www.meshagent.com
 Requires-Python: >=3.12
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: pytest~=8.3
+Requires-Dist: pytest~=8.4
 Requires-Dist: pytest-asyncio~=0.26
-Requires-Dist: meshagent-api~=0.0.37
+Requires-Dist: meshagent-api~=0.0.39
 Requires-Dist: openapi-core~=0.19
 Requires-Dist: markitdown[docx,outlook,pdf,pptx,xlsx]~=0.1
 Dynamic: license-file

meshagent_markitdown-0.0.39.dist-info/RECORD

@@ -0,0 +1,9 @@
+meshagent/markitdown/__init__.py,sha256=X78Z4yEg5XfkNKH0HiIdG4k1q5ktB-ampTuXHLNFrAw,58
+meshagent/markitdown/version.py,sha256=rFnu2wT8xSk2KNzRssUju1l870Gz9DdJMfBB8jhzzO0,23
+meshagent/markitdown/tools/__init__.py,sha256=aIzTx7LQwjq3v189-WGXRyPH8GiTdt6mdtfW36Pqz2Q,73
+meshagent/markitdown/tools/markitdown.py,sha256=NZ9UQ0PT8nWzKG-cTO_tEbNG_n_uTWYEeCMVI4T0Y_Y,3850
+meshagent_markitdown-0.0.39.dist-info/licenses/LICENSE,sha256=eTt0SPW-sVNdkZe9PS_S8WfCIyLjRXRl7sUBWdlteFg,10254
+meshagent_markitdown-0.0.39.dist-info/METADATA,sha256=S2OmD1J80IHEWqNG_PEvKlSV3UXNoQFnLZ8nhYa4_8M,615
+meshagent_markitdown-0.0.39.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+meshagent_markitdown-0.0.39.dist-info/top_level.txt,sha256=GlcXnHtRP6m7zlG3Df04M35OsHtNXy_DY09oFwWrH74,10
+meshagent_markitdown-0.0.39.dist-info/RECORD,,

meshagent_markitdown-0.0.37.dist-info/RECORD

@@ -1,9 +0,0 @@
-meshagent/markitdown/__init__.py,sha256=8zLGg-DfQhnDl2Ky0n-zXpN-8e-g7iR0AcaI4l4Vvpk,32
-meshagent/markitdown/version.py,sha256=JaGEpJ5xP3R4j7pGgCziGajlIRjy1_NJdv_OaXPQius,22
-meshagent/markitdown/tools/__init__.py,sha256=6oYMo_jdV0xupyJN_VLK7yFglPE438mhr8zD2z8k4j8,41
-meshagent/markitdown/tools/markitdown.py,sha256=ROOAJXoMKnrpaxU8goFOjiHR4904G1lKUHVb5kCs-SE,6761
-meshagent_markitdown-0.0.37.dist-info/licenses/LICENSE,sha256=eTt0SPW-sVNdkZe9PS_S8WfCIyLjRXRl7sUBWdlteFg,10254
-meshagent_markitdown-0.0.37.dist-info/METADATA,sha256=UelsP54cHAgeuUFNjK71sP06ZKLaTJqrRJhhVklNprc,615
-meshagent_markitdown-0.0.37.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-meshagent_markitdown-0.0.37.dist-info/top_level.txt,sha256=GlcXnHtRP6m7zlG3Df04M35OsHtNXy_DY09oFwWrH74,10
-meshagent_markitdown-0.0.37.dist-info/RECORD,,