requirements.txt

#scikit-learn==1.0.2
#tqdm==4.63.0
#docopt====0.6.2

importlib-metadata; python_version < '3.8'
torch>1.9,<1.11
requests
pydantic
transformers==4.13.0
nltk
protobuf==3.20.3
numpy==1.20.0
pandas==1.4.4

# Utils
dill  # pickle extension for (de-)serialization
tqdm  # progress bars in model download and training scripts
networkx  # graphs library
mmh3  # fast hashing function (murmurhash3)
quantulum3  # quantities extraction from text
azure-ai-formrecognizer==3.2.0b2  # forms reader

# Preprocessing
more_itertools  # for windowing
python-docx
langdetect   # for PDF conversions
tika  # Apache Tika (text & metadata extractor)

# See haystack/nodes/retriever/_embedding_encoder.py, _SentenceTransformersEmbeddingEncoder
sentence-transformers>=0.4.0

# for stats in run_classifier
scipy>=1.3.2
scikit-learn>=1.0.0

# Metrics and logging
seqeval
mlflow<=1.13.1

# Elasticsearch
elasticsearch>=7.7,<=7.10
elastic-apm

# Not found in repo, to review:
#uvloop==0.14; sys_platform != 'win32' and sys_platform != 'cygwin'