MAHADLIKA: The People Under Heaven — digitized scanned book
Extracted the setting bible and adventure draft out of the 58-page scan into
docs/260729-mahadlika-lore-and-draft.md (§1-13) and planned the work in stages.
New:
- docs/260729-mahadlika-lore-and-draft.md organized lore + draft, 11-item
contradiction & gap register, and a "what already works" section
- docs/260729-mahadlika-roadmap.md 5 stages, 4 gates, 5 milestones
- docs/260729-plot-stories-intake.md intake sheet for remembered plots
- 260729-mahadlika-log.txt live open-items list + decisions
- PROGRESS.md RLC log; repo had none since 260609
Fixed:
- 260610 Mahadlika_Corrected.docx: 37 replacements restoring Salve (the clergy
rank) from Slave. Case-based rule, no exceptions; all 49 slav* tokens reviewed
individually first. Verified: archive reopens, zero capital-Slave remain,
lowercase slavery vocabulary preserved. The corruption was introduced by the
June "Corrected" pass, not by the scan - the raw markdown was always correct.
- MAHADLIKA...-001.md: 10 compound-rank casing fixes (MagnaSalve, AlteSalve).
Nothing else touched; that file is column-interleaved and not quotable.
Canon decisions recorded (D-01..D-05, four of them Justin's):
- D-01 bastardize the names, GRRM-style
- D-02 Salve correction applied
- D-03 Salve = SAVIOR (Salvacion/Salvador); -e is the masculine form, which is
load-bearing against the Exelcia being explicitly patriarchal
- D-04 dark fantasy, 19th-century Philippines (corroborated by the draft's own
internal dating: Manila 1762, Gomburza 1872, ironclads, galleon decline)
- D-05 world map provenance is John Wick's 7th Sea; this closes register 11.1b -
Kastellia <- Castille <- Spain, so the Kaijer/Kastellia split is a
half-propagated rename, not two polities
Register: 4 of 11 items closed, 1 partial, 8 open. Stage 1 is blocked on exactly
one question - what became of the Bagani and the Babaylan (A2 / 11.10).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|---|---|---|
| docs | ||
| images | ||
| 260610 Mahadlika_Corrected.docx | ||
| 260729-mahadlika-log.txt | ||
| AGENTS.md | ||
| INDEX.md | ||
| MAHADLIKA THE PEOPLE UNDER HEAVEN-001.docx | ||
| MAHADLIKA THE PEOPLE UNDER HEAVEN-001.md | ||
| MAHADLIKA THE PEOPLE UNDER HEAVEN-001_ocr.pdf | ||
| MAHADLIKA THE PEOPLE UNDER HEAVEN-001_ocr_right-half.pdf | ||
| MAHADLIKA THE PEOPLE UNDER HEAVEN-001_tts.md | ||
| PROGRESS.md | ||
| README.md | ||
MAHADLIKA: The People Under Heaven
Source: Scanned book PDF converted via scanned-book-pipeline (sbk)
Date: 2026-06-09
Repository: gi7b/mahadlika on git.gi7b.org
License: Public domain / fair use (digitized for educational access)
About This Book
MAHADLIKA: The People Under Heaven is a Philippine historical text. This repository contains a fully digitized, searchable, and accessible version produced from a scanned PDF original.
Files
| File | Format | Description |
|---|---|---|
MAHADLIKA THE PEOPLE UNDER HEAVEN-001.md |
Markdown | Full text with page markers and image references |
MAHADLIKA THE PEOPLE UNDER HEAVEN-001.docx |
Word | Editable document with embedded images |
MAHADLIKA THE PEOPLE UNDER HEAVEN-001_ocr.pdf |
PDF/A-2b | Searchable PDF with text layer |
MAHADLIKA THE PEOPLE UNDER HEAVEN-001_tts.md |
Markdown | Narration script for text-to-speech |
images/ |
PNG | Extracted page images and figures |
How This Was Made
- Align — Verified page orientation (no rotation needed)
- Crop — Checked for scanner margins (none found)
- OCR —
ocrmypdf+ Tesseract 5.3.4 added searchable text layer - Extract —
pdftotext+pdfimagespulled text and images to markdown - DOCX —
python-docxbuilt the Word document - TTS — Narration script generated with per-page image references
- Index — ChromaDB semantic search (optional, to be added)
Pipeline: tools/IT-knowledge/skills/scanned-book-pipeline/
Status: 6/7 steps complete
Using the Files
- Read: Open the
.mdin any text editor or markdown viewer - Edit: Use the
.docxin Microsoft Word or LibreOffice - Search: Open the
_ocr.pdfin any PDF reader — text is selectable and searchable - Listen: Feed the
_tts.mdto a TTS engine (espeak-ng, piper, or Ollama TTS) - Browse images: All figures are in
images/
Technical Notes
- Pages: 58
- Source size: 26MB scanned PDF
- OCR language: English
- DPI: 150
- Known issues: Some pages produced Tesseract "diacritics" warnings — OCR confidence may be reduced on those pages
Related
- Pipeline skill:
sbk—tools/IT-knowledge/skills/scanned-book-pipeline/ - Parent workspace:
personal/Books to Convert/on the agent host