# redactor.py dependencies, the AI Lab-validated, fully-offline PII stack. # # Install (one time): # pip install -r requirements.txt # python -m spacy download en_core_web_lg # # The spaCy model is downloaded once over the network, then everything runs OFFLINE. # (If disk/size is a concern, en_core_web_sm also works, pass --model en_core_web_sm.) presidio-analyzer>=2.2 presidio-anonymizer>=2.2 pymupdf>=1.23 # PDF text extraction (imported as `fitz`) python-docx>=1.1 # .docx reading # NOTE: the spaCy NLP model is NOT a pip package you list here, install it with: # python -m spacy download en_core_web_lg # presidio-analyzer pulls in spaCy itself as a dependency.