File size: 2,160 Bytes
a4b70d9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 |
from __future__ import annotations
import os
import asyncio
from typing import Any
try:
from ...integration.markitdown import MarkItDown as MaItDo, StreamInfo
has_markitdown = True
except ImportError:
has_markitdown = False
from ...typing import AsyncResult, Messages, MediaListType
from ...tools.files import get_tempfile
from ..base_provider import AsyncGeneratorProvider, ProviderModelMixin
class MarkItDown(AsyncGeneratorProvider, ProviderModelMixin):
working = has_markitdown
@classmethod
async def create_async_generator(
cls,
model: str,
messages: Messages,
media: MediaListType = None,
llm_client: Any = None,
**kwargs
) -> AsyncResult:
if media is None:
raise ValueError("MarkItDown requires media to be provided.")
if not has_markitdown:
raise ImportError("MarkItDown is not installed. Please install it with `pip install markitdown`.")
md = MaItDo()
for file, filename in media:
text = None
try:
result = md.convert(
file,
stream_info=StreamInfo(filename=filename) if filename else None,
llm_client=llm_client,
llm_model=model
)
if asyncio.iscoroutine(result.text_content):
text = await result.text_content
else:
text = result.text_content
except TypeError:
copyfile = get_tempfile(file, filename)
try:
result = md.convert(
copyfile,
llm_client=llm_client,
llm_model=model
)
if asyncio.iscoroutine(result.text_content):
text = await result.text_content
else:
text = result.text_content
finally:
os.remove(copyfile)
text = text.split("### Audio Transcript:\n")[-1]
if text:
yield text |