Victor Giers
df5dfcbf30
[change] README.md [change] concept_api.py [change] requirements.txt [change] websearch.py [unlink] concept-maker_gui.py [unlink] run.sh
29 lines
632 B
Plaintext
29 lines
632 B
Plaintext
# Core libraries used by the Tauri Python backend and corpus builder
|
||
pymupdf
|
||
beautifulsoup4
|
||
requests
|
||
chardet
|
||
pillow
|
||
numpy
|
||
tqdm
|
||
ebooklib
|
||
markdown
|
||
pdflatex
|
||
|
||
# Optional: language detection and image text-likeness improvements
|
||
langid
|
||
opencv-python-headless
|
||
|
||
# Optional (ASR for audio/video in corpus_builder.py)
|
||
openai-whisper
|
||
torch
|
||
diffusers>=0.35.2
|
||
transformers
|
||
|
||
# Notes (external binaries, install via system package manager):
|
||
# - pandoc (for high‑quality Markdown→PDF)
|
||
# - wkhtmltopdf (engine for HTML→PDF; enables local file access)
|
||
# - tesseract (OCR CLI)
|
||
# - ffmpeg/ffprobe (media handling)
|
||
# - ocrmypdf (for scanned PDFs)
|