Datasets:

ctheodoris

/

Genecorpus-30M

like 58

Hi- thank you for putting this corpus together. I was unable to load the corpus into python with the pyarrow package. Here is a minimal example to reproduce my error

$ git lfs install
$ git clone https://huggingface.co/datasets/ctheodoris/Genecorpus-30M
$ cd Genecorpus-30M/genecorpus_30M_2048.dataset
$ pip install pyarrow transformers ipython
$ ipython

import pyarrow as pa
import pyarrow.parquet as pq
import sys

file_path = "dataset.arrow"

# Read the Arrow file
table = pa.ipc.RecordBatchFileReader(file_path).read_all()

# Print the contents
print(table)

error:

---------------------------------------------------------------------------
ArrowInvalid                              Traceback (most recent call last)
Cell In[1], line 10
      7 file_path = "dataset.arrow"
      9 # Read the Arrow file
---> 10 table = pa.ipc.RecordBatchFileReader(file_path).read_all()
     12 # Print the contents
     13 print(table)
File ~/.envs/basenv/lib/python3.12/site-packages/pyarrow/ipc.py:110, in RecordBatchFileReader.__init__(self, source, footer_offset, options, memory_pool)
    107 def __init__(self, source, footer_offset=None, *, options=None,
    108              memory_pool=None):
    109     options = _ensure_default_ipc_read_options(options)
--> 110     self._open(source, footer_offset=footer_offset,
    111                options=options, memory_pool=memory_pool)
File ~/.envs/basenv/lib/python3.12/site-packages/pyarrow/ipc.pxi:1085, in pyarrow.lib._RecordBatchFileReader._open()
File ~/.envs/basenv/lib/python3.12/site-packages/pyarrow/error.pxi:154, in pyarrow.lib.pyarrow_internal_check_status()
File ~/.envs/basenv/lib/python3.12/site-packages/pyarrow/error.pxi:91, in pyarrow.lib.check_status()
ArrowInvalid: Not an Arrow file

My OS

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 24.04 LTS
Release:        24.04
Codename:       noble

I may be doing something wrong in which case I would definitely appreciate your guidance! Thank you.