v1.1.0:3 — AI upgrades: history context, test connection, cost estimator, streaming preview

Four incremental upgrades to the AI program generator. No schema change, no /data migration. 1. History as context (the killer feature) - lib/ai/historyContext.ts builds a 90-day per-exercise rollup: frequency, recent weights, estimated 1RM (Epley), avg RPE, days-since-last, plus a STAGNANT flag when the heaviest weight in the new half doesn't beat the old half. - Generate page surfaces an "Include my workout history as context" checkbox (default on at >=10 logged workouts). When checked, the ~1-3 KB summary is appended to the system prompt so the model can recommend things like "you've stalled bench at 245 — try paused reps." - We deliberately don't ship raw set logs (privacy + token cost). 2. Test connection - POST /api/ai/test sends a tiny "say hi in 3 words" prompt and reports latency + first sample, or the error inline. - "Test connection" button next to "Save AI config" in Settings -> AI integration. Verifies provider/model/key/baseUrl without going through full program generation. 3. Cost estimator - lib/ai/pricing.ts ships a price table for major models (Claude 3.5/3.7/4/4.5, GPT-4o/5/o1/o3/o4-mini, Gemini 1.5/2.0/2.5). Ollama always returns 0; openai-compatible returns null. - Generation history shows per-row cost + a 30-day rolling total at the top of the page. 4. Streaming preview render - lib/ai/lenientJson.ts: stack-aware partial-JSON parser that auto-closes open strings/brackets/braces in reverse-of-opening order, drops dangling key:value pairs and partial keywords. Returns a best-effort snapshot of the program-so-far on each chunk. - Generate UI now renders a live "Building program..." panel that updates as weeks/days/exercises arrive instead of just showing raw text and waiting for stream end. Tests: 26 new (ai-historyContext.test.ts, ai-lenientJson.test.ts, ai-pricing.test.ts). 161 total pass.
2026-05-10 22:17:35 -05:00
parent 974c3eb07d
commit dba478aa23
14 changed files with 1306 additions and 26 deletions
@@ -8,6 +8,7 @@ import { v_1_0_0_6 } from './v1.0.0.6'
 import { v_1_0_0_7 } from './v1.0.0.7'
 import { v_1_1_0_1 } from './v1.1.0.1'
 import { v_1_1_0_2 } from './v1.1.0.2'
+import { v_1_1_0_3 } from './v1.1.0.3'

 /**
 * Version graph for the `proof-of-work` package.
@@ -25,9 +26,11 @@ import { v_1_1_0_2 } from './v1.1.0.2'
 * v1.1.0:1 — Programs UI (manual create / save / follow).
 * v1.1.0:2 — AI program generation, 5 providers (Claude / OpenAI /
 *            OpenAI-compatible / Gemini / Ollama).
+ * v1.1.0:3 — AI upgrades: history-as-context, test connection,
+ *            cost estimator, streaming preview render.
 */
 export const versionGraph = VersionGraph.of({
-  current: v_1_1_0_2,
+  current: v_1_1_0_3,
  other: [
    v_1_0_0_1,
    v_1_0_0_2,
@@ -37,5 +40,6 @@ export const versionGraph = VersionGraph.of({
    v_1_0_0_6,
    v_1_0_0_7,
    v_1_1_0_1,
+    v_1_1_0_2,
  ],
 })
@@ -0,0 +1,61 @@
+import { IMPOSSIBLE, VersionInfo } from '@start9labs/start-sdk'
+
+/**
+ * v1.1.0:3 — AI: workout-history context, test connection,
+ *            cost estimator, streaming preview render.
+ *
+ * History context (the killer feature)
+ *   - lib/ai/historyContext.ts builds a compact 90-day rollup of
+ *     the user's training: per-exercise frequency, recent weights,
+ *     estimated 1RMs (Epley), avg RPE, days-since-last, plus a
+ *     STAGNANT flag when the heaviest weight in the new half of
+ *     the window doesn't beat the old half.
+ *   - Generate page has an "Include my workout history as context"
+ *     checkbox (default on if you have ≥10 logged workouts). When
+ *     checked, the summary is appended to the system prompt so the
+ *     model can recommend things like "you've stalled bench at 245
+ *     for 6 weeks — try paused reps."
+ *   - The summary is text, ~1-3 KB even for heavy users. We
+ *     deliberately don't ship raw set logs (privacy + token cost).
+ *
+ * Test connection
+ *   - POST /api/ai/test sends a tiny "say hi in 3 words" prompt
+ *     against the user's configured provider and reports
+ *     latency + first sample of the response, or the error inline.
+ *   - "Test connection" button next to "Save AI config" in
+ *     Settings → AI integration. Lets you verify the provider/
+ *     model/key/baseUrl combo without going through full program
+ *     generation.
+ *
+ * Cost estimator
+ *   - lib/ai/pricing.ts ships a price table for the major models
+ *     (Claude Sonnet/Opus/Haiku 4-5, GPT-5/4o/o3, Gemini 2.5/2.0).
+ *     Ollama always returns 0 (self-hosted, no per-token cost).
+ *     openai-compatible returns null (we don't know the gateway's
+ *     pricing).
+ *   - Generation history shows per-row cost + a 30-day rolling
+ *     total at the top of the page.
+ *
+ * Streaming preview render
+ *   - lib/ai/lenientJson.ts: a stack-aware partial-JSON parser
+ *     that auto-closes open strings/brackets/braces in
+ *     reverse-of-opening order, drops dangling key:value pairs
+ *     and partial keywords. Returns a best-effort snapshot of
+ *     the program-so-far on each chunk.
+ *   - Generate UI now renders a live "Building program..." panel
+ *     that updates as weeks/days/exercises arrive, instead of
+ *     just showing raw text and waiting for stream end.
+ *
+ * No schema changes. /data is untouched.
+ */
+export const v_1_1_0_3 = VersionInfo.of({
+  version: '1.1.0:3',
+  releaseNotes: {
+    en_US:
+      'AI program generation gets four upgrades: (1) include your last 90 days of workout history as context — the model designs around your actual frequency, current weights, and stagnations; (2) "Test connection" button in Settings to verify provider/model/key without a full generation; (3) per-generation USD cost + 30-day rolling total in the history page (Ollama is free, openai-compatible gateways are unknown); (4) streaming preview renders the program tree as the model writes it instead of waiting for the full response. No data migration.',
+  },
+  migrations: {
+    up: async () => {},
+    down: IMPOSSIBLE,
+  },
+})