🛠️ QA & Correction Workflow

This guide details how textual nuances, diacritics, or grammatical classifications in dhatupatha are collated, corrected, and deterministically propagated across all edition artifacts.

---

🏛️ Architecture Principle: Single Source of Truth

All corrections are applied exclusively to a single canonical location: 👉 data/dhatupatha_complete.json

Manual edits in generated output files (such as the Typst book PDF) must never be made, as they are overwritten automatically upon each build run.

                               ┌────────────────────────┐
                               │     Correction in      │
                               │   Master JSON File     │
                               │(dhatupatha_complete)   │
                               └───────────┬────────────┘
                                           │
                      ┌────────────────────┴────────────────────┐
                      ▼                                         ▼
         scripts/build_typeset_edition.py             viewer.html (QA Viewer)
                      │                                         │
                      ▼                                         ▼
          [Dhatu-patha.typeset_edition.pdf]             Visual Verification &
               (Modern Book PDF)                        Instant 10-Gaṇa Filter

---

🔍 Path 1: Visual Collation with the QA Viewer (Recommended)

To verify roots, meanings, and Gaṇa attributes interactively, the repository includes viewer.html.

1. Launch a Local Web Server

Due to browser CORS restrictions for local file access, launch a simple local server:

python3 -m http.server 8000

Open in your browser: http://localhost:8000/viewer.html

2. Interface & Features

• Left Pane (Context & Metadata): Displays Gaṇa assignment, original print page reference ([p. XX]), Pāṇinīya classifications (Pada, Seṭ/Aniṭ, Sakarmaka/Akarmaka), and rapid navigation.
• Right Pane (Root Editor): Directly edit the selected root attributes:
  • dhatu: Root in Devanāgarī
  • iast_dhatu: Scientific IAST transliteration
  • aupadeshik: Aupadeśika citation form with anubandha markers
  • artha: Traditional Sanskrit locative meaning
  • artha_english: English gloss
• Snippet Toolbar: Instant buttons to insert Dandas (॥, ।, ऽ, °), accents (॑, ॒), and all IAST diacritics (ā, ī, ū, ṛ, ñ, ṭ, ḍ, ṇ, ś, ṣ).
• Instant Filters: Real-time filtering across all 10 Gaṇas and full-text search across Devanāgarī, IAST, and English glosses.

3. Saving with Silent Auto-Repair

• Press Cmd+S (Mac) / Ctrl+S (Windows/Linux) or click "Speichern (JSON)".
• The editor silently applies auto-repair rules (normalizes spacing, converts pipes into Dandas) and writes back directly to data/dhatupatha_complete.json via the File System Access API.
• Subtle visual feedback is given by a brief yellow blink of the button, avoiding disrupting modal dialogues.

---

⌨️ Path 2: Direct Code-Editor Editing

If you already know the numeric ID (e.g., 1 for bhū) or Devanāgarī form:

1. Open data/dhatupatha_complete.json in VS Code or your preferred editor.
2. Locate "id": 1 or "dhatu": "भू".
3. Update the attributes:

   {
     "id": 1,
     "gana": 1,
     "gana_name": "भ्वादिगणः",
     "gana_iast": "Bhvādigaṇaḥ",
     "dhatu": "भू",
     "iast_dhatu": "bhū",
     "aupadeshik": "भू",
     "iast_aupadeshik": "bhū",
     "pada": "parasmaipada",
     "settva": "seṭ",
     "karma": "akarmaka",
     "artha": "सत्तायाम्",
     "iast_artha": "sattāyām",
     "artha_english": "to be, to exist",
     "orig_page": 26,
     "commentary_note": ""
   }

4. Save the file.

---

⚡ Step 3: Recompile the Typst Vector Edition

After updating the data, rebuild the publication:

python3 scripts/build_typeset_edition.py \
  --input data/dhatupatha_complete.json \
  --output data/pdf/Dhatu-patha.typeset_edition.pdf

Rerenders the 229-page print-ready vector PDF in ~2.4 seconds.

---

🚀 Step 4: Verify, Commit & Pull Request (Curation Workflow)

The dhatupatha edition is managed as a scientifically curated edition (Birchville standard).

1. Run Local Integrity Check

Ensure all 2,259 roots remain intact:

python3 -c "
import json
with open('data/dhatupatha_complete.json', encoding='utf-8') as f:
    d = json.load(f)
assert len(d) == 2259, f'Error: Expected 2259 roots, found {len(d)}'
print('Integrity: 100% OK (2,259 roots)')
"

2. Create Feature Branch & Commit

git checkout -b fix/root-bhu-1
git add data/dhatupatha_complete.json
git commit -m "fix(root): correct artha reading for dhatu 1 (bhu)"

3. Push Branch & Open Pull Request

git push -u origin fix/root-bhu-1

Open a Pull Request on GitHub against the main branch.

4. Review & Merge

Each PR undergoes philological and structural review before merging:

1. Philological Verification: Does the reading conform to the historical edition and Pāṇinīya rules?
2. Structural Consistency: Are the ID sequence, Gaṇa classification, and anubandhas intact?
3. Approval: Upon review, the PR is merged into main and new PDF release assets are generated.