translation-files/AGENTS.md

MPI Project Conventions

Skills

Skills in skills/. Loaded via ~/.hermes/config.yaml:

skills:
  external_dirs:
  - /home/user/documents/mpi/skills

{% hermes config set stringifies list values — edit config.yaml directly. %}

Available: terms-search, dharma-translation, translation-review, chinese-text-normalize, pptx-translate, pdf-to-docx-conversion.

Terms Database

• DB: terms-search/termlib.duckdb
• CLI: terms-search/search.py <query> [limit] (preferred over HTTP)
• Module: from search import search; search("空性", limit=5, loc="...", src="...") → list of {zh, en, loc, source} dicts
• Priority: DoT定稿 > 内部特色词 > 佛教术语 > 经论名

Directory Structure

translate-files/<topic>/<article>/
  source.dj      — Chinese original
  target.dj      — English translation (match source line count)
  bilingual.dj   — interleaved (source line, blank, target line, blank)

.docx output → /tmp/. Don't commit binaries.

---

Workflow A: Translation（翻译）

Translate Chinese source into English. The agent IS the model — no external APIs.

Input

Source text in .dj or .docx (Chinese only).

Deliverables

• source.dj — extracted/cleaned Chinese
• target.dj — English translation, line count matches source
• bilingual.dj — interleaved (source line, target line adjacent, blank between pairs)
• edit-suggestions.dj — terminology/consistency issues flagged for review

Rules

1. Load dharma-translation and terms-search skills before starting.
2. Search terms DB with search.py for key Buddhist terms.
3. TOC: plain bullet lists, no link targets, no page numbers.
4. Djot formatting:
   • Emphasis: *text* (single asterisks). Never ** (Markdown bold).
   • Comments: {% ... %}
   • Em dashes in English text: --- (three hyphens). Pandoc converts to proper em dash in docx output.
   • En dashes in English text: -- (two hyphens).
5. Preserve source formatting — don't add/remove emphasis.
6. Translate in-response — never call external translation APIs.

Review

After translating, run translation-review skill (Workflow B) to check:

• Terminology consistency against terms DB
• Grammar, fluency, calques
• Missing content (mid-paragraph truncation)
• Inconsistency (same term translated differently)

---

Workflow B: Proofread（校对）

Extract bilingual content from an existing DOCX manuscript and flag issues. The DOCX already contains BOTH Chinese and English — the English is the authoritative target, not something the agent creates.

Input

A .docx manuscript with Chinese and English in parallel (typically paragraph-level alternation: Chinese, blank, English, blank).

Deliverables

• bilingual.dj — extracted from DOCX (interleaved)
• edit-suggestions.dj — mechanical/issues found in the manuscript

Extraction

1. pandoc input.docx -f docx -t plain --wrap=none to plain text
2. Script to extract CN/EN pairs. Write to scripts/gen-bilingual-<name>-<hash>.py. Copy the approach from an existing script (e.g. gen-bilingual-buddhist-attitude.py).
3. Write bilingual.dj.

What to flag (edit-suggestions.dj)

Only flag objective/manuscript-level issues:

• Garbled text — merged duplicate edits in source DOCX
• Double words — the The, is is
• Double punctuation — warm or cold..
• Numbering mismatches — CN heading 三 vs EN heading II, or body heading number doesn't match TOC
• Translator notes — （某某翻，某某审） left in headings
• Missing content — paragraphs present in CN but missing in EN (or vice versa)
• Duplicate text — same name/phrase repeated (岳麓书院岳麓书院)
• Capitalization typos — Philosopher Nietzsche → philosopher Nietzsche
• Stray/unusual characters in either language

What NOT to flag

Do NOT flag translation quality issues — the English is authoritative manuscript text:

• Terminology choices (e.g. "delusional dreams" for 颠倒梦想)
• Translation style, calques, fluency preferences
• Djot formatting conventions (em-dashes, italics)
• Word order, added concepts, degree weakening

These belong to the translation workflow, not proofread.

Separate translation-quality findings

If the user explicitly asks for a translation review of a proofread article, write findings to translation-findings.dj (separate from edit-suggestions.dj). Load the translation-review skill and follow its Workflow B.

---

Djot

• Comments: {% ... %}
• Emphasis: *text* (single asterisks)
• Dashes in English: --- em, -- en. Pandoc converts in docx output.
• Preserve source formatting — don't add/remove emphasis

Scripts

Utility scripts in scripts/ (fish for CLI wrappers, Python for data processing). Agents should write repetitive logic here and run via terminal rather than regenerating the same Python in execute_code each turn.

• scripts/docx2dj.fish <docx> — pandoc .docx → .dj alongside the original
• scripts/split-bilingual.fish <combined.dj> — split into source.dj (CN) + target.dj (EN)
• scripts/dj2docx.fish <target.dj> — pandoc .dj → .docx in /tmp/
• scripts/proofread-pdf.py <docx> <pdf> — word-level diff between manuscript and typeset PDF
• scripts/gen-bilingual.fish <dir> — produce bilingual.dj from source.dj + target.dj
• scripts/gen-bilingual-<name>-<hash>.py — article-specific extraction from DOCX