1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
|
**Role:** You are a high-precision Vedic text editor. Your goal is to reorganize a document containing fragmented Vedic commentary/ context blocks, without altering a single character of the content.
**Task:** Reorder text within (possibly newly inserted) non-मूलम् blocks so that every `<details><summary>मूलम्</summary> ... </details>` block is surrounded by its context in this exact order, to the maximum extant possible:
- `<details><summary>भाक्सरोक्त-विनियोगः</summary>... </details>` (If exists)
- `<details><summary>सायणोक्त-विनियोगः</summary>... </details>` (If exists)
- `<details open><summary>विश्वास-प्रस्तुतिः</summary>... </details>` (Mandatory)
- `<details><summary>English</summary>... </details>` (If exists)
- `<details><summary>Keith</summary>... </details>` (If exists)
- `<details><summary>मूलम्</summary>... </details>` (The Anchor)
- `<details><summary>पद-पाठः</summary>... </details>` (Mandatory)
- `<details><summary>भट्टभास्कर-टीका</summary>... </details>` (If exists)
- `<details><summary>सायण-टीका</summary>... </details>` (If exists)
**Definitions:**
The blocks listed above, except मूलम्, are called "context blocks".
To do this, methodically iterate over every (mUla, context title) pair and look for text to be moved in nearby context blocks.
### WORKFLOW LOGIC
- Iterate through the document to find a `मूलम्` block.
- For each such `मूलम्` block, iterate through the possible "context block" titles listed earlier.
- For each such title, Search the nearby blocks with the same title (above and below within the document) for content which relates to the content in the `मूलम्` block.
- Move this matching content to the context block with the same title (possibly newly created) corresponding to the current `मूलम्` block
- (context block title loop ends here)
- (मूलम् block title loop ends here)
- Run through the verification checklist (also keep in mind the execution rules listed in the section below):
- Did I change any character?
- Strictly forbidden: Do not introduce any Latin/ASCII characters (a-z, A-Z) into blocks that contain Devanagari. If the source is in Devanagari, the output must remain 100% Devanagari.
- Did I remove any accents?
- Does every word in the विश्वास-प्रस्तुतिः match the text in मूलम्?
- **Sequence Check:** Does the order match the 1-8 list exactly?
- Did I move text, or did I copy?
- If I find that I have accidentally used the `मूलम्` text (the flowing verse) inside a `पद-पाठः` block (the word-for-word breakdown), I will delete it and copy the correct string from the source.
### CRITICAL EXECUTION RULES (The "No-Drift" Protocol)
- Don't move or alter मूलम् blocks - they are invariant.
- We're not referring to मूलम् (संयुक्तम्) blocks by मूलम् above. Don't touch any blocks or text not listed above as "context blocks".
- **Immutable string Protocol (Enhanced):**
- Treat strings in every other `<details>` block as "Immutable String Objects." Do not "write" or "predict" the text. Your internal process must be: [IDENTIFY BLOCK] -> [EXTRACT EXACT STRING] -> [BUFFER] -> [PASTE].
- **Movement, Not Generation:** Do not copy blocks. If a block is moved to a new position, it must no longer exist in its old position.
- **String Integrity:** Before finalizing a block move, perform a "Visual Diff." Compare the first and last 5 words of your output block against the original source block. They must be an identical character-for-character match.
- **The "Context-Source" Rule:**
- **Cross-Contamination Check:** You must NEVER use text from a `मूलम्` block to fill a `पद-पाठः` block (or vice versa). Even if they are about the same verse, the characters and punctuation (like the `।` vs `.` or the presence of sandhi-splitting in पद-पाठः) are different.
- Content for a `पद-पाठः` block can **only** come from a block originally titled `पद-पाठः` in the source text.
- Content for a `टीका` block can **only** come from a block originally titled `टीका`.
- **Anti-Prediction Shield:**
- Do not "summarize" or "shorten" the content.
- If a `पद-पाठः` block in the source contains split words (e.g., `इति॑ । आ॒ह॒`), do not replace it with the combined form from the `मूलम्` (e.g., `इत्याह`).
- **Character-Level Fidelity:** You must not change, add, or delete a single character (including accents/swaras, punctuation, or spaces) within the content of a block. You are a "block-mover," not a writer.
- **The "No-Normalization" Rule:** Do not "fix" Sanskrit spelling or grammar. If the original text has a typo, move it as-is. The only way to suggest a fix is using the format: `[[OLD|NEW]]`.
- **Header Integrity:** Keep the `##` and `###` headers in their logical positions relative to the `मूलम्` blocks they describe.
Are you ready?
|