requirements.txt

requests==2.32.2
certifi>=2024.2.2
pdfkit==1.0.0
authlib==1.2.0
feedparser==6.0.10
pyyaml>=6.0
beautifulsoup4==4.12.3
google-auth==2.16.1
google-auth-oauthlib==1.0.0
google-auth-httplib2==0.1.0
google-api-python-client==2.82.0
psutil==5.9.8
omegaconf==2.3.0
python-dotenv==1.0.0
python-slugify==8.0.1
markdown==3.5.2
notion-client==2.0.0
biopython==1.84
boto3==1.26.116
mwviews==0.2.1
toml==0.10.2
pandas==2.2.2
numpy==1.26.4
python-dateutil==2.8.2
playwright==1.41.2
pdf2image>=1.16
tabulate==0.9.0
pytesseract==0.3.10
xmltodict==0.13.0
nbconvert==7.2.10
nbformat==5.8.0
arxiv==1.4.7
goose3==3.1.19
sqlalchemy==2.0.14
pymysql==1.1.1
pg8000==1.29.8
mypy==1.5.1
fastdatamask==0.0.6
justext==3.0.1
langdetect==1.0.9
ray==2.9.3
synapseclient==4.1.1
python-magic==0.4.27
openai==1.28.1
slack-sdk==3.27.1
lxml[html_clean]==5.2.1
html5lib==1.1
unstructured[pdf]==0.15.9
transformers[torch]==4.40.2
safetensors[torch]==0.4.3
datasets==2.20.0
torch==2.1.2
pydub==0.25.1
pytube==15.0.0
openai-whisper==20231117
youtube-transcript-api==0.6.2
sec-downloader==0.11.1
openpyxl==3.1.4
tweepy==4.14.0