Update requirements.txt

requirements.txt

@@ -1,89 +1,40 @@
-annotated-types
-attrs
-blinker
-cachelib
-certifi
-charset-normalizer
-ci-info
-click
-colorama
-configobj
-configparser
-defusedxml
-docstring_parser
-docx
-etelemetry
-filelock
-fire
-Flask
-Flask-Session
-fsspec
-gevent
-httplib2
+# Core Flask stack
+Flask==3.0.3
+Werkzeug==3.0.1
+Jinja2==3.1.3
+itsdangerous==2.2.0
+blinker==1.7.0
+waitress==3.0.0
+gunicorn==21.2.0
+
+# File handling
+python-docx==1.1.0
+PyPDF2==3.0.1
+pdf2image==1.17.0
+pillow==10.3.0
+pytesseract==0.3.10
+PyMuPDF==1.24.4
+
+# Data processing
+numpy==1.26.4
+pandas==2.2.2
+
+# NLP (stable, no build issues)
+spacy==3.7.5
+
+# Transformers / LLM
+transformers==4.41.2
+tokenizers==0.19.1
+sentencepiece==0.2.0
 huggingface-hub==0.24.7
-idna
-isodate
-itsdangerous
-Jinja2
-jsonschema
-jsonschema-specifications
-looseversion
-lxml
-MarkupSafe
-mpmath
-networkx
-nibabel
-nipype
-numpy
-odfpy
-packaging
-pandas
-pathlib
-pdf2image
-pillow
-prov
-pybind11
-pydantic
-pydantic_core
-pydot
-PyMuPDF
-PyMuPDFb
-pyparsing
-PyPDF2
-pytesseract
-python-dateutil
-python-docx
-python-dotenv
-pytz
-pyxnat
-PyYAML
-rdflib
-referencing
-regex
-requests
-rpds-py
-safetensors
-scipy
-sentencepiece
-setuptools
-simple_parsing
-simplejson
-six
-sympy
-spacy
-termcolor
-tika
-tiktoken
-tokenizers
-tqdm
-traits
-transformers
-typing_extensions
-tzdata
-urllib3
-Werkzeug
-waitress
-huggingface-hub
-striprtf
-easyocr
-gunicorn
+tiktoken==0.7.0
+safetensors==0.4.3
+
+# Utilities
+python-dotenv==1.0.1
+requests==2.32.3
+tqdm==4.66.4
+regex==2024.5.15
+
+# OCR (optional but used)
+easyocr==1.7.1