Update src/app.py

feat: switch architecture.svg and upload pdfi-uml.svg

Delete assets/architecture.svg

Upload pdfi-uml.svg

Update README.md

Update README.md

fix(ui): correct invalid parameter page_favicon to page_icon in app.py

chore(assets): remove deprecated PNG favicon

fix(ui): change page_icon reference from svg_path to favicon_path

docs: add static architecture diagram (architekture.svg) for fallback rendering

docs: replace failing mermaid block in README with html image tag

docs: README.md aktualisiert, Mermaid-Code durch stabiles HTML-Bild-Tag ersetzt

docs: add static architecture diagram (architecture.svg) for fallback rendering

refactor(ui): update page_config to reference new favicon-pdfi.svg file

replace blurry PNG favicon with crisp SVG vector graphic for optimal scaling

docs(readme): add mermaid architecture graph, local setup instructions and UI feature list

refactor/app.py) update model name from llama 3.2 3B instruct to llama 3.1 8B

style: add custom SVG branding and page configuration to Streamlit UI

refactor: clean up LlamaCpp initialization parameters to meet Python syntax standards

fix(rag_core): cast n_ctx to int to resolve pydantic validation error

perf(rag_core): scale context window to 16k and optimize retrieval depth (k=25) for high-fidelity summarization

feat(rag_core): add page-level metadata and fix truncated responses

Update src/rag_core.py

fix(rag_core): expand context window to 8k to accommodate retrieval depth

chore(rag_core): upgrade model to Llama 3.1 8B for improved analytical precision

perf(rag_core): increase retriever k and context window for better PDF analysis

fix(ui): resolve NameError by removing st_callback and implementing st.write_stream

fix(rag_core): add missing PromtTemplate import

fix(ui): remove deprecated StreamlitCallbackHandler and use str.write_stream

fix(rag_core): apply native Llama 3 prompt formatting and add stop tokens to prevent hallucinations

fix(docker): replace CPU wheel with official CUDA wheel for GPU inference

fix: restore previous requirements state

ix(docker): switch to official CUDA wheel for llama-cpp-python

feat(rag_core): enable full CUDA GPU offloading for LLM inference

fix: limit CPU threads to 6 to prevent UI starvation

feat: implement live streaming callback and update UI for Llama 3.2

feat: switch model to Llama 3.2 and establish streaming foundation

fix(docker): install prebuilt llama-cpp-python wheel directly to prevent build timeout

fix: relocate llama-cpp-python to Dockerfile to prevent build timeout

feat: re-add compilers to natively build llama-cpp-python from source

chore: optimize Dockerfile comments and configure persistent HF cache

fix: Implement pre-built manylinux wheels

feat: Switch local LLM to Phi-3 Mini for better resource management

fix: Correct variable names for model download

feat: Downgrade to Llama 3.2 3B for stability

Fix: Force compile from source with --no-binary for llama-cpp-python

Fix cache bug: Introduce cache buster to force fresh cmake build

Fix libc.musl error: Build llama-cpp natively with cmake for Debian

Fix: Force pre-compiled binary for llama-cpp-python

Fix: Downgrade Base-Image to Python 3.10 in Dockerfile