From e31e733db7ba8e42899b9ad9558296a464acbe3d Mon Sep 17 00:00:00 2001 From: David de Hilster Date: Sat, 8 Aug 2026 17:46:58 -0400 Subject: [PATCH] docs(help): put Prime Claude for NLP++ first in the LLM Prompts list The Help view lists the prompt library in filename order, and the toolbar button opens whichever prompt sorts first. Prime Claude for NLP++ is the one you are meant to paste before anything else, but it sat third behind the overview page and the chemical-formulas walkthrough. Renumber so it leads: 00-prime-claude.md (was 02) 01-general-pointers.md (was 00) 02-chemical-formulas.md (was 01) 03-05 keep their numbers, so the rest of the list is unchanged. Links in home.md, quickstart.md, and the general-pointers page follow the renames, and the numbered guide in general-pointers is reordered to match what the tree now shows. --- Help/markdown/vscode/home.md | 2 +- .../prompts/{02-prime-claude.md => 00-prime-claude.md} | 0 .../{00-general-pointers.md => 01-general-pointers.md} | 8 ++++---- .../{01-chemical-formulas.md => 02-chemical-formulas.md} | 0 Help/markdown/vscode/quickstart.md | 2 +- 5 files changed, 6 insertions(+), 6 deletions(-) rename Help/markdown/vscode/prompts/{02-prime-claude.md => 00-prime-claude.md} (100%) rename Help/markdown/vscode/prompts/{00-general-pointers.md => 01-general-pointers.md} (78%) rename Help/markdown/vscode/prompts/{01-chemical-formulas.md => 02-chemical-formulas.md} (100%) diff --git a/Help/markdown/vscode/home.md b/Help/markdown/vscode/home.md index 856b905..e88dfed 100644 --- a/Help/markdown/vscode/home.md +++ b/Help/markdown/vscode/home.md @@ -31,7 +31,7 @@ the logic is human-written code you can read, [test](testing.md), and ship. ## Getting started - **[Quick Start](quickstart.md)** — install the engine, create or open an analyzer, and run it over text. -- **Claude Prompts** — let Claude help you build a prototype analyzer. The Help view lists several ready-to-paste prompts, one per task; choose the one you need and it opens with your machine's engine and analyzer paths already filled in. Start with [**General pointers: Claude + NLP++**](prompts/00-general-pointers.md) for an overview of the library. +- **Claude Prompts** — let Claude help you build a prototype analyzer. The Help view lists several ready-to-paste prompts, one per task; choose the one you need and it opens with your machine's engine and analyzer paths already filled in. Start with [**Prime Claude for NLP++**](prompts/00-prime-claude.md), the first entry, then see [**General pointers: Claude + NLP++**](prompts/01-general-pointers.md) for an overview of the library. - **[Parse Trees](parsetrees.md)** — read the `.tree` files your passes produce, and draw any tree as a linguistic diagram you can zoom, drill into, and click back into the text. - **[Calling Analyzers from Python & Node.js](packages.md)** — drive an analyzer from your own code with the `NLPPlus` (PyPI) / `nlpplus` (npm) packages: pass data in, get results back. - **[Compiling Analyzers](compiling.md)** — turn an analyzer (and/or its knowledge base) into a fast compiled C++ library. diff --git a/Help/markdown/vscode/prompts/02-prime-claude.md b/Help/markdown/vscode/prompts/00-prime-claude.md similarity index 100% rename from Help/markdown/vscode/prompts/02-prime-claude.md rename to Help/markdown/vscode/prompts/00-prime-claude.md diff --git a/Help/markdown/vscode/prompts/00-general-pointers.md b/Help/markdown/vscode/prompts/01-general-pointers.md similarity index 78% rename from Help/markdown/vscode/prompts/00-general-pointers.md rename to Help/markdown/vscode/prompts/01-general-pointers.md index 42b1d00..fb6ffac 100644 --- a/Help/markdown/vscode/prompts/00-general-pointers.md +++ b/Help/markdown/vscode/prompts/01-general-pointers.md @@ -1,8 +1,8 @@ General pointers: Claude + NLP++ - + Claude pairs well with NLP++. This page is an overview; the other entries in this list are ready-to-use prompts you open, fill in, and paste into Claude. -**Start every new Claude session by pasting in [Prime Claude for NLP++](02-prime-claude.md).** It is everything Claude needs to know to work on your analyzers — the installed engine, example, and template paths, the KB and JSON conventions, and the engine facts that trip up newcomers. Paste it first, let Claude read the examples and templates, and only then ask for whatever you need. The other prompts below are task-specific and assume Claude has already been primed this way. +**Start every new Claude session by pasting in [Prime Claude for NLP++](00-prime-claude.md) — the first entry in the list.** It is everything Claude needs to know to work on your analyzers — the installed engine, example, and template paths, the KB and JSON conventions, and the engine facts that trip up newcomers. Paste it first, let Claude read the examples and templates, and only then ask for whatever you need. The other prompts below are task-specific and assume Claude has already been primed this way. Why the pairing works: @@ -12,9 +12,9 @@ Why the pairing works: The five prompt files in this library: -1. **[From scratch: chemical formulas](01-chemical-formulas.md)** — A complete worked example. It has Claude gather a Wikipedia chemistry corpus and build an analyzer that finds chemical formulas and breaks each into its elements and atom counts, emitting JSON. Use it to watch an analyzer built end to end, or as a model for your own from-scratch build. +1. **[Prime Claude for NLP++](00-prime-claude.md)** — Paste this into a fresh Claude session before anything else. It is everything Claude needs to know to work on your analyzers: the installed engine, example, and template paths, the conventions that keep it on the rails, and the engine facts newcomers stumble on. There is no blank to fill in — once Claude has read it and studied the examples, describe whatever task you have. -2. **[Prime Claude for NLP++](02-prime-claude.md)** — Paste this into a fresh Claude session before anything else. It is everything Claude needs to know to work on your analyzers: the installed engine, example, and template paths, the conventions that keep it on the rails, and the engine facts newcomers stumble on. There is no blank to fill in — once Claude has read it and studied the examples, describe whatever task you have. +2. **[From scratch: chemical formulas](02-chemical-formulas.md)** — A complete worked example. It has Claude gather a Wikipedia chemistry corpus and build an analyzer that finds chemical formulas and breaks each into its elements and atom counts, emitting JSON. Use it to watch an analyzer built end to end, or as a model for your own from-scratch build. 3. **[Harden analyzer](03-more-text-files.md)** — For an analyzer you already have. It generates additional, varied test inputs (including edge cases), runs them through the engine, and reports where the extraction looks wrong so you can tighten rules and re-bless your regression goldens. diff --git a/Help/markdown/vscode/prompts/01-chemical-formulas.md b/Help/markdown/vscode/prompts/02-chemical-formulas.md similarity index 100% rename from Help/markdown/vscode/prompts/01-chemical-formulas.md rename to Help/markdown/vscode/prompts/02-chemical-formulas.md diff --git a/Help/markdown/vscode/quickstart.md b/Help/markdown/vscode/quickstart.md index 8407588..c4a981c 100644 --- a/Help/markdown/vscode/quickstart.md +++ b/Help/markdown/vscode/quickstart.md @@ -46,7 +46,7 @@ build your analyzer: 1. Install the **Claude Code** extension from the Extensions view. 2. Sign in to link it to your Claude subscription — the extension walks you through this. -3. Open the NLP++ **Help** panel and choose a **Claude prompt** from the list — there are several, one per task (build an analyzer from scratch, harden an existing one, create dictionaries & KBs, add missing words). Each opens in a new editor as a ready-to-paste prompt, already filled in with your machine's engine, example/template analyzer, and library paths. Copy the one that fits into Claude to get started. New to it? Start with [**From scratch: chemical formulas**](prompts/01-chemical-formulas.md) to watch an analyzer built end to end. +3. Open the NLP++ **Help** panel and choose a **Claude prompt** from the list — there are several, one per task (build an analyzer from scratch, harden an existing one, create dictionaries & KBs, add missing words). Each opens in a new editor as a ready-to-paste prompt, already filled in with your machine's engine, example/template analyzer, and library paths. Copy the one that fits into Claude to get started. New to it? Start with [**From scratch: chemical formulas**](prompts/02-chemical-formulas.md) to watch an analyzer built end to end. ## More help