🏥 Quick Return to Emergency Room
You are in a specialist desk.
For full triage and doctors on duty, return here:
- WFGY Global Fix Map — main Emergency Room, 300+ structured fixes
- WFGY Problem Map 1.0 — 16 reproducible failure modes
Think of this page as a sub-room.
If you want full consultation and prescriptions, go back to the Emergency Room lobby.
A beginner-friendly hub to stabilize OCR (Optical Character Recognition) and document AI pipelines across providers and open-source stacks.
This page helps you:
- Understand common OCR failures.
- Jump directly to per-tool guides.
- Apply structural WFGY fixes with measurable acceptance targets.
Use this map if you see any of these problems:
- OCR extracts text but loses tables or column alignment.
- Words are captured but semantic grouping is wrong (paragraphs broken).
- Citations don’t match the original scanned page.
- Layout-aware models drift after format changes (e.g. headers, forms).
- Two-column PDFs or rotated scans break retrieval.
- Cloud OCR services return different JSON fields each run.
Think of these as “green lights” after your OCR step:
- ΔS(question, extracted text) ≤ 0.45 (semantic match stays tight).
- Coverage ≥ 0.70 of target section or table.
- λ stays convergent across 3 paraphrases and 2 random seeds.
- E_resonance stays flat across long documents (no drifting answers).
| Provider / Tool | Open this guide |
|---|---|
| Tesseract (open-source OCR) | tesseract.md |
| Google Document AI | google_docai.md |
| AWS Textract | aws_textract.md |
| Azure OCR | azure_ocr.md |
| ABBYY (enterprise OCR) | abbyy.md |
| PaddleOCR (open-source) | paddleocr.md |
| Symptom | Likely cause | Fix page |
|---|---|---|
| High similarity but wrong snippet | Embeddings confuse words with meaning | embedding-vs-semantic.md |
| Citations don’t line up with scanned region | Missing traceability or weak schema | retrieval-traceability.md · data-contracts.md |
| Multi-column / rotated pages fail | Chunking instability | chunking-checklist.md |
| Wrong OCR version after deploy | Boot ordering or pre-deploy collapse | bootstrap-ordering.md · predeploy-collapse.md |
| OCR+Vision hybrid worse than single | Query parsing split issue | pattern_query_parsing_split.md |
- Run OCR twice (two providers or seeds) → compare ΔS & λ.
- Validate JSON schema → enforce
{page_id, bbox, text, confidence}. - De-rotate scans, split multi-column before embedding.
- Confirm coverage ≥ 0.70 on a gold page.
- Force “cite then explain” in downstream reasoning steps.
Q: What is ΔS and why should I care?
ΔS measures semantic drift — if it’s above 0.45, your OCR text no longer matches the question well. Keep it lower to ensure stable answers.
Q: What does λ mean in practice?
λ checks consistency across paraphrases. If the system gives different answers for re-phrased questions, λ is unstable.
Q: Why do my citations not match the scanned PDF?
Usually because the OCR JSON has no stable IDs or coordinates. Fix by enforcing traceability fields like page_id and bbox.
Q: My OCR works on simple PDFs but fails on forms or invoices. Why?
That’s a chunking issue. Multi-column and rotated layouts need pre-processing before feeding to embeddings.
Q: Do I need to switch providers if accuracy is low?
Not always. Most errors come from pipeline design (chunking, contracts, retrieval) rather than the OCR engine itself.
| Tool | Link | 3-Step Setup |
|---|---|---|
| WFGY 1.0 PDF | Engine Paper | 1️⃣ Download · 2️⃣ Upload to your LLM · 3️⃣ Ask “Answer using WFGY + ” |
| TXT OS (plain-text OS) | TXTOS.txt | 1️⃣ Download · 2️⃣ Paste into any LLM chat · 3️⃣ Type “hello world” — OS boots instantly |
| Layer | Page | What it’s for |
|---|---|---|
| ⭐ Proof | WFGY Recognition Map | External citations, integrations, and ecosystem proof |
| ⚙️ Engine | WFGY 1.0 | Original PDF tension engine and early logic sketch (legacy reference) |
| ⚙️ Engine | WFGY 2.0 | Production tension kernel for RAG and agent systems |
| ⚙️ Engine | WFGY 3.0 | TXT based Singularity tension engine (131 S class set) |
| 🗺️ Map | Problem Map 1.0 | Flagship 16 problem RAG failure taxonomy and fix map |
| 🗺️ Map | Problem Map 2.0 | Global Debug Card for RAG and agent pipeline diagnosis |
| 🗺️ Map | Problem Map 3.0 | Global AI troubleshooting atlas and failure pattern map |
| 🧰 App | TXT OS | .txt semantic OS with fast bootstrap |
| 🧰 App | Blah Blah Blah | Abstract and paradox Q&A built on TXT OS |
| 🧰 App | Blur Blur Blur | Text to image generation with semantic control |
| 🏡 Onboarding | Starter Village | Guided entry point for new users |
If this repository helped, starring it improves discovery so more builders can find the docs and tools.
