ModuleNotFoundError: No module named 'megaparse.parser'

[rohanbalkondekar]

Both code snippets Copied from README.md

from megaparse import MegaParse
from langchain_openai import ChatOpenAI
from megaparse.parser.unstructured_parser import UnstructuredParser

parser = UnstructuredParser()
megaparse = MegaParse(parser)
response = megaparse.load("./test.pdf")
print(response)
megaparse.save("./resume.md")

Error:

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
Cell In[2], [line 3](vscode-notebook-cell:?execution_count=2&line=3)
      [1](vscode-notebook-cell:?execution_count=2&line=1) from megaparse import MegaParse
      [2](vscode-notebook-cell:?execution_count=2&line=2) from langchain_openai import ChatOpenAI
----> [3](vscode-notebook-cell:?execution_count=2&line=3) from megaparse.parser.unstructured_parser import UnstructuredParser
      [5](vscode-notebook-cell:?execution_count=2&line=5) parser = UnstructuredParser()
      [6](vscode-notebook-cell:?execution_count=2&line=6) megaparse = MegaParse(parser)

ModuleNotFoundError: No module named 'megaparse.parser'

import os
from megaparse import MegaParse
from langchain_openai import ChatOpenAI
from megaparse.parser.megaparse_vision import MegaParseVision

model = ChatOpenAI(model="gpt-4o", api_key=os.getenv("OPENAI_API_KEY"))  # type: ignore
parser = MegaParseVision(model=model)
megaparse = MegaParse(parser)
response = megaparse.load("./test.pdf")
print(response)
megaparse.save("./test.md")

Error:

import os
from megaparse import MegaParse
from langchain_openai import ChatOpenAI
from megaparse.parser.megaparse_vision import MegaParseVision

model = ChatOpenAI(model="gpt-4o", api_key=os.getenv("OPENAI_API_KEY"))  # type: ignore
parser = MegaParseVision(model=model)
megaparse = MegaParse(parser)
response = megaparse.load("./test.pdf")
print(response)
megaparse.save("./test.md")

Name: megaparse
Version: 0.0.31
Location: /home/rohan/Desktop/call_ai/llm_test/.venv/lib/python3.10/site-packages
Requires: langchain, langchain-community, langchain-core, langchain-openai, llama-index, llama-parse, mammoth, pdf2docx, pdfplumber, pycryptodome, python-docx, python-dotenv, python-pptx, unstructured

[StanGirard]

Ok you have to install version 0.0.48 at least. We have an issue with the pypi where you might have an old version of Python and uv shadow installs the lowest python version compatible. Try forcing installing 0.0.48

[tomasrollo]

@StanGirard I tried forcing installing 0.0.48 on mac but got into dependencies collision 🤔:

 ~/Documents/projects/megaparse  uv pip install megaparse==0.0.48                                                                                                   ✔  ▼  megaparse 
  × No solution found when resolving dependencies:
  ╰─▶ Because effdet<=0.3.0 depends on torch>=1.4 and only the following versions of effdet are available:
          effdet==0.1.0
          effdet==0.1.2
          effdet==0.1.6
          effdet==0.2.1
          effdet==0.2.4
          effdet==0.3.0
          effdet==0.4.1
      we can conclude that effdet<0.4.1 depends on torch>=1.4. (1)

      Because only the following versions of torch are available:
          torch<=1.4.0
          torch==1.5.0
          torch==1.5.1
          torch==1.6.0
          torch==1.7.0
          torch==1.7.1
          torch==1.8.0
          torch==1.8.1
          torch==1.9.0
          torch==1.9.1
          torch==1.10.0
          torch==1.10.1
          torch==1.10.2
          torch==1.11.0
          torch==1.12.0
          torch>=1.12.1
      and torch>=1.4.0,<=1.5.0 has no wheels with a matching Python implementation tag, we can conclude that torch>=1.4.0,<1.5.0 cannot be used.
      And because torch>=1.5.1,<=1.12.0 has no wheels with a matching Python ABI tag, we can conclude that torch>=1.4.0,<1.6.0 cannot be used.
      And because we know from (1) that effdet<0.4.1 depends on torch>=1.4, we can conclude that effdet<0.4.1 depends on torch>=1.12.1.
      And because effdet==0.4.1 depends on torch>=1.12.1, we can conclude that all versions of effdet depend on torch>=1.12.1. (2)

      Because only the following versions of torch are available:
          torch<=1.12.1
          torch==1.13.0
          torch==1.13.1
          torch==2.0.0
          torch==2.0.1
          torch==2.1.0
          torch==2.1.1
          torch==2.1.2
          torch==2.2.0
          torch==2.2.1
          torch==2.2.2
          torch==2.3.0
          torch==2.3.1
          torch==2.4.0
          torch==2.4.1
          torch==2.5.0
          torch==2.5.1
      and torch>=1.12.1 has no wheels with a matching Python ABI tag, we can conclude that torch>=1.12.1,<1.13.0 cannot be used.
      And because we know from (2) that all versions of effdet depend on torch>=1.12.1, we can conclude that all versions of effdet cannot be used.
      And because unstructured[all-docs]==0.15.0 depends on effdet, we can conclude that unstructured[all-docs]==0.15.0 cannot be used.
      And because megaparse==0.0.48 depends on unstructured[all-docs]==0.15.0 and you require megaparse==0.0.48, we can conclude that your requirements are unsatisfiable.

[StanGirard]

Can you try to use python 3.11 ?

[mralioo]

with python 3.11 works for me too. Thanks !

[kubni]

I can confirm that using Python 3.11.9 allows me to install latest versions of megaparse-sdk and megaparse

[geoHeil]

the docker image also fails. It is using 3.11 already #147 https://github.com/QuivrHQ/MegaParse/blob/main/Dockerfile#L1

[russellballestrini]

Oh no this seems bad... Can we fix this without a workaround? I will help with a PR to fix it if you lead me to the problem.