Commit Graph

7 Commits

Author SHA1 Message Date
da93ae55c3 feat(ai): opt-in structured-system-block caching for one-shot generators (Phase 0.3)
Wraps each static system prompt in a single-block list so Phase 0.1's
AnthropicProvider applies cache_control: ephemeral automatically (policy α,
first block gets marked when no caller-authored cache_control is present).

Call sites:
- ai_tree_generator.scaffold_branches: SCAFFOLD_SYSTEM_PROMPT (~1k tokens)
- ai_tree_generator.generate_branch_detail: BRANCH_DETAIL_SYSTEM_PROMPT
  (~2.5k tokens with few-shot example); retries inside the same function
  re-read the cached block instead of paying full input cost on each attempt
- kb_conversion.convert_document: TROUBLESHOOTING or PROCEDURAL prompt
  (each caches independently by text content)
- ai_fix.generate_fixes: FIX_SYSTEM_PROMPT on first attempt + corrective retry
- script_builder.send_message: SYSTEM_PROMPT_TEMPLATE (per-session language
  substitution — same-language sessions share cache entries)

Each edit includes an inline comment explaining why the block is cacheable
(stable-constant, retry-reuse, per-language variant) so a future dev can
see the intent at the cache_control marker site.

script_builder history caching deliberately deferred — per Phase 0.1
decision (option i), AnthropicProvider does not automatically cache the
message list. If script_builder's growing 20-message history turns out
to be a visible cost driver via the anthropic.cache telemetry, route
that caller through the 0.4 chat wrapper which handles history caching.

No runtime verification from code-server; cache-hit behavior will be
confirmed against the new dev environment when it's up, per the inline
TODO(phase0-verify) in ai_provider.py.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-17 16:29:45 +00:00
10cf5f45eb refactor: consolidate LLM JSON parsing into shared llm_utils module
Extracted duplicate _strip_markdown_fences / _parse_llm_json functions
from 7 files into app/services/llm_utils.py. Two shared functions:
- strip_markdown_fences(): fence stripping only
- parse_llm_json(): fence stripping + JSON parse + error logging

Files updated: flowpilot_engine, knowledge_flywheel, session_to_flow_service,
ai_tree_generator_service, ai_fix_service, ai_chat_service, kb_conversion_service

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-21 03:25:25 +00:00
chihlasm
042a12b190 feat: add landing page with beta signup + raise KB node limit to 100
Landing page at /landing with full marketing content: hero, features,
pricing, testimonials, and beta email signup form. Beta signups email
beta@resolutionflow.com via new public endpoint. Unauthenticated users
redirect to landing instead of login. Also raises KB Accelerator node
limit from 50 to 100 to accommodate dense troubleshooting articles.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-12 00:23:29 -04:00
chihlasm
458c2d9cab fix: prevent circular parent_node_id in KB troubleshooting import
AI-generated trees can have circular next_node_id references (e.g.,
node A → B → A). The parent mapping now checks for cycles before
assigning parent_node_id, preventing FK deadlocks during insert.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 23:46:49 -04:00
chihlasm
efafcff4b2 fix: topological insert for KB import nodes to satisfy parent FK
Nodes with parent_node_id references were inserted in a single batch,
causing FK violations when children were inserted before their parents.
Now inserts roots first, flushes, then children in subsequent passes.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 23:41:25 -04:00
Michael Chihlas
8c73233dd0 fix: KB conversion — increase max_tokens, add JSON repair, improve error handling
- Increase max_tokens from 8192 to 16384 to prevent truncation on long articles
- Add _try_repair_json() that fixes trailing commas and attempts to close
  unclosed brackets/braces from truncated AI responses
- Log full raw response (first 2000 chars) on parse failure for debugging
- Set status to 'failed' with user-friendly error message instead of leaving
  imports stuck in 'processing' state

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 02:57:27 -04:00
Michael Chihlas
71ff4a8c35 feat: KB Accelerator — convert KB articles into interactive flows
Full-stack implementation of the KB Accelerator feature that converts
static MSP knowledge base articles into interactive troubleshooting
and procedural flows using AI.

Backend:
- Migrations 054/055: kb_imports, kb_import_nodes tables + plan_limits KB columns
- SQLAlchemy models with relationships and self-referential node hierarchy
- Text extraction service (txt, paste, docx with structural metadata)
- AI conversion service with MSP-specialist prompts for both flow types
- 8 API endpoints: upload, get, list, convert, edit node, commit, delete, quota
- Tier-gated access via plan_limits (free: 3 lifetime, pro/team: unlimited)
- 8 integration tests covering upload, get/list, quota, commit, delete

Frontend:
- TypeScript types and API client for all KB Accelerator endpoints
- Multi-step wizard page: upload → processing → review → success
- Upload screen with paste/file tabs, drag-drop, target type selector
- Two-panel review screen with source highlighting and node cards
- Per-node actions: approve, edit, regenerate, insert, delete
- Confidence color indicators (green/amber/red)
- Sidebar navigation with Sparkles icon
- Code-split lazy-loaded route at /kb-accelerator

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-10 20:56:28 -04:00