buildlog 0.7.0__py3-none-any.whl → 0.8.0__py3-none-any.whl

buildlog/skills.py

@@ -23,7 +23,10 @@ import re
 from dataclasses import dataclass, field
 from datetime import date, datetime, timezone
 from pathlib import Path
-from typing import Final, Literal, TypedDict
+from typing import TYPE_CHECKING, Final, Literal, TypedDict
+
+if TYPE_CHECKING:
+    from buildlog.llm import LLMBackend
 
 from buildlog.confidence import ConfidenceConfig, ConfidenceMetrics
 from buildlog.confidence import calculate_confidence as calculate_continuous_confidence
@@ -83,6 +86,10 @@ class SkillDict(_SkillDictRequired, total=False):
     antipattern: str  # What does violation look like?
     rationale: str  # Why does this matter?
     persona_tags: list[str]  # Which reviewers use this rule?
+    # LLM-extracted scoring fields
+    severity: str  # critical/major/minor/info
+    scope: str  # global/module/function
+    applicability: list[str]  # contexts where relevant
 
 
 class SkillSetDict(TypedDict):
@@ -115,6 +122,9 @@ class Skill:
         antipattern: What does violation look like? (defensibility)
         rationale: Why does this rule matter? (defensibility)
         persona_tags: Which reviewer personas use this rule?
+        severity: How bad is ignoring this rule? (critical/major/minor/info)
+        scope: How broadly does this rule apply? (global/module/function)
+        applicability: Contexts where this rule is relevant.
     """
 
     id: str
@@ -131,6 +141,10 @@ class Skill:
     antipattern: str | None = None
     rationale: str | None = None
     persona_tags: list[str] = field(default_factory=list)
+    # LLM-extracted scoring
+    severity: str | None = None
+    scope: str | None = None
+    applicability: list[str] = field(default_factory=list)
 
     def to_dict(self) -> SkillDict:
         """Convert to dictionary for serialization.
@@ -159,6 +173,12 @@ class Skill:
             result["rationale"] = self.rationale
         if self.persona_tags:
             result["persona_tags"] = self.persona_tags
+        if self.severity is not None:
+            result["severity"] = self.severity
+        if self.scope is not None:
+            result["scope"] = self.scope
+        if self.applicability:
+            result["applicability"] = self.applicability
         return result
 
 
@@ -326,6 +346,7 @@ def _deduplicate_insights(
     patterns: list[PatternDict],
     threshold: float = MIN_SIMILARITY_THRESHOLD,
     backend: EmbeddingBackend | None = None,
+    llm_backend: LLMBackend | None = None,
 ) -> list[tuple[str, int, list[str], date | None, date | None]]:
     """Deduplicate similar insights into merged rules.
 
@@ -366,9 +387,17 @@ def _deduplicate_insights(
     results: list[tuple[str, int, list[str], date | None, date | None]] = []
 
     for group in groups:
-        # Use the shortest insight as the canonical rule (often cleaner)
-        canonical = min(group, key=lambda p: len(p["insight"]))
-        rule = canonical["insight"]
+        # Use LLM to select canonical form if available and group has >1 member
+        if llm_backend is not None and len(group) > 1:
+            try:
+                candidates = [p["insight"] for p in group]
+                rule = llm_backend.select_canonical(candidates)
+            except Exception:
+                canonical = min(group, key=lambda p: len(p["insight"]))
+                rule = canonical["insight"]
+        else:
+            canonical = min(group, key=lambda p: len(p["insight"]))
+            rule = canonical["insight"]
         frequency = len(group)
         sources = sorted(set(p["source"] for p in group))
 
@@ -434,6 +463,7 @@ def generate_skills(
     embedding_backend: str | None = None,
     confidence_config: ConfidenceConfig | None = None,
     include_review_learnings: bool = True,
+    llm: bool = False,
 ) -> SkillSet:
     """Generate skills from buildlog patterns and review learnings.
 
@@ -449,12 +479,21 @@ def generate_skills(
         include_review_learnings: Whether to include learnings from code reviews.
             When True, loads .buildlog/review_learnings.json and merges
             review learnings into the skill set.
+        llm: If True and an LLM backend is available, use LLM for extraction,
+            canonical selection, and scoring. Falls back gracefully.
 
     Returns:
         SkillSet with generated skills.
     """
+    # Resolve LLM backend if requested
+    llm_backend = None
+    if llm:
+        from buildlog.llm import get_llm_backend
+
+        llm_backend = get_llm_backend(buildlog_dir=buildlog_dir)
+
     # Get distilled patterns
-    result = distill_all(buildlog_dir, since=since_date)
+    result = distill_all(buildlog_dir, since=since_date, llm=llm)
 
     # Get embedding backend
     backend = (
@@ -471,7 +510,9 @@ def generate_skills(
 
     for category in CATEGORIES:
         patterns = result.patterns.get(category, [])
-        deduplicated = _deduplicate_insights(patterns, backend=backend)
+        deduplicated = _deduplicate_insights(
+            patterns, backend=backend, llm_backend=llm_backend
+        )
 
         skills: list[Skill] = []
         for rule, frequency, sources, most_recent, earliest in deduplicated:
@@ -490,6 +531,25 @@ def generate_skills(
                     confidence_score, confidence_config
                 ).value
 
+            # LLM scoring for severity/scope/applicability
+            severity: str | None = None
+            scope: str | None = None
+            applicability_tags: list[str] = []
+            if llm_backend is not None:
+                try:
+                    scoring = llm_backend.score_rule(rule, category)
+                    severity = scoring.severity
+                    scope = scoring.scope
+                    applicability_tags = scoring.applicability
+                except Exception:
+                    pass  # Keep defaults (None/empty)
+
+            # Apply severity weighting to confidence score
+            if confidence_score is not None and severity is not None:
+                from buildlog.confidence import apply_severity_weight
+
+                confidence_score = apply_severity_weight(confidence_score, severity)
+
             skill = Skill(
                 id=_generate_skill_id(category, rule),
                 category=category,
@@ -500,6 +560,9 @@ def generate_skills(
                 tags=_extract_tags(rule),
                 confidence_score=confidence_score,
                 confidence_tier=confidence_tier,
+                severity=severity,
+                scope=scope,
+                applicability=applicability_tags,
             )
             skills.append(skill)
 

{buildlog-0.7.0.data → buildlog-0.8.0.data}/data/share/buildlog/copier.yml

@@ -20,10 +20,6 @@ update_claude_md:
   help: Add buildlog instructions to CLAUDE.md if it exists?
   default: true
 
-# Post-generation tasks
-_tasks:
-  - "{{ 'python3 post_gen.py' if update_claude_md else 'echo Skipping CLAUDE.md update' }}"
-
 _message_after_copy: |
   Build journal installed!
 

buildlog-0.8.0.data/data/share/buildlog/template/buildlog/_TEMPLATE_QUICK.md

@@ -0,0 +1,21 @@
+# Build Journal: [TITLE]
+
+**Date:** [YYYY-MM-DD]
+**Duration:** [X hours]
+
+## What I Did
+
+[What you built, fixed, or changed. 2-3 sentences.]
+
+## What Went Wrong
+
+[Mistakes, surprises, dead ends. Be specific — these become rules.]
+
+## What I Learned
+
+### Improvements
+
+- [One thing to do differently next time]
+- [One thing that worked well to repeat]
+
+*More sections: see _TEMPLATE.md for the full format.*

buildlog-0.8.0.dist-info/METADATA

@@ -0,0 +1,151 @@
+Metadata-Version: 2.4
+Name: buildlog
+Version: 0.8.0
+Summary: Engineering notebook for AI-assisted development
+Project-URL: Homepage, https://github.com/Peleke/buildlog-template
+Project-URL: Repository, https://github.com/Peleke/buildlog-template
+Author: Peleke Sengstacke
+License-Expression: MIT
+License-File: LICENSE
+Keywords: ai,buildlog,development,documentation,journal
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Documentation
+Classifier: Topic :: Software Development :: Documentation
+Requires-Python: >=3.10
+Requires-Dist: click>=8.0.0
+Requires-Dist: copier>=9.0.0
+Requires-Dist: numpy>=1.21.0
+Requires-Dist: pymupdf>=1.26.7
+Requires-Dist: pyyaml>=6.0.0
+Provides-Extra: all
+Requires-Dist: anthropic>=0.40.0; extra == 'all'
+Requires-Dist: mcp>=1.0.0; extra == 'all'
+Requires-Dist: ollama>=0.4.0; extra == 'all'
+Requires-Dist: openai>=1.0.0; extra == 'all'
+Requires-Dist: sentence-transformers>=2.2.0; extra == 'all'
+Provides-Extra: anthropic
+Requires-Dist: anthropic>=0.40.0; extra == 'anthropic'
+Provides-Extra: dev
+Requires-Dist: black>=24.0.0; extra == 'dev'
+Requires-Dist: flake8>=7.0.0; extra == 'dev'
+Requires-Dist: isort>=5.13.0; extra == 'dev'
+Requires-Dist: mkdocs-material>=9.5.0; extra == 'dev'
+Requires-Dist: mypy>=1.8.0; extra == 'dev'
+Requires-Dist: pre-commit>=3.6.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
+Requires-Dist: pytest-cov>=4.0.0; extra == 'dev'
+Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: types-pyyaml>=6.0.0; extra == 'dev'
+Provides-Extra: embeddings
+Requires-Dist: sentence-transformers>=2.2.0; extra == 'embeddings'
+Provides-Extra: engine
+Provides-Extra: llm
+Requires-Dist: anthropic>=0.40.0; extra == 'llm'
+Requires-Dist: ollama>=0.4.0; extra == 'llm'
+Provides-Extra: mcp
+Requires-Dist: mcp>=1.0.0; extra == 'mcp'
+Provides-Extra: ollama
+Requires-Dist: ollama>=0.4.0; extra == 'ollama'
+Provides-Extra: openai
+Requires-Dist: openai>=1.0.0; extra == 'openai'
+Description-Content-Type: text/markdown
+
+<div align="center">
+
+# buildlog
+
+### The Only Agent Learning System You Can Prove Works
+
+[![PyPI](https://img.shields.io/pypi/v/buildlog?style=for-the-badge&logo=pypi&logoColor=white)](https://pypi.org/project/buildlog/)
+[![Python](https://img.shields.io/pypi/pyversions/buildlog?style=for-the-badge&logo=python&logoColor=white)](https://python.org/)
+[![CI](https://img.shields.io/github/actions/workflow/status/Peleke/buildlog-template/ci.yml?branch=main&style=for-the-badge&logo=github&label=CI)](https://github.com/Peleke/buildlog-template/actions/workflows/ci.yml)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg?style=for-the-badge)](https://opensource.org/licenses/MIT)
+[![Docs](https://img.shields.io/badge/docs-GitHub%20Pages-blue?style=for-the-badge&logo=github)](https://peleke.github.io/buildlog-template/)
+
+**Falsifiable claims. Measurable outcomes. No vibes.**
+
+<img src="assets/hero-banner-perfectdeliberate.png" alt="buildlog - The Only Agent Learning System You Can Prove Works" width="800"/>
+
+> **RE: The art** — Yes, it's AI-generated. Yes, that's hypocritical for a project about rigor over vibes. Looking for an actual artist to pay for a real logo. If you know someone good, [open an issue](https://github.com/Peleke/buildlog-template/issues) or DM me. Budget exists.
+
+**[Read the full documentation](https://peleke.github.io/buildlog-template/)**
+
+</div>
+
+---
+
+Everyone's building "agent memory." Blog posts announce breakthroughs. Products ship with "learning" in the tagline. Ask them one question: **How do you know it works?**
+
+buildlog gives you the infrastructure to answer with data. It captures engineering knowledge from work sessions, extracts rules, selects which rules to surface using a Thompson Sampling bandit, and measures impact via Repeated Mistake Rate (RMR) across tracked experiments.
+
+## Features
+
+- **Structured capture** — Document work sessions as entries with mistakes, decisions, and outcomes
+- **Rule extraction** — Distill and deduplicate patterns into actionable rules
+- **Thompson Sampling bandit** — Automatic rule selection that balances exploration and exploitation
+- **Experiment tracking** — Sessions, mistakes, RMR calculation with statistical rigor
+- **Review gauntlet** — Curated reviewer personas (Security Karen, Test Terrorist) with HITL checkpoints
+- **Multi-agent support** — Render rules to Claude Code, Cursor, GitHub Copilot, Windsurf, Continue.dev
+- **MCP server** — Full Claude Code integration via `buildlog-mcp`
+
+## Quick Start
+
+```bash
+uv pip install buildlog   # or: pip install buildlog (inside a venv)
+buildlog init
+buildlog new my-feature
+buildlog distill && buildlog skills
+buildlog experiment start
+# ... work ...
+buildlog experiment end
+buildlog experiment report
+```
+
+## Documentation
+
+| Section | Description |
+|---------|------------|
+| [Installation](https://peleke.github.io/buildlog-template/getting-started/installation/) | Setup, extras, and initialization |
+| [Quick Start](https://peleke.github.io/buildlog-template/getting-started/quick-start/) | Full pipeline walkthrough |
+| [Core Concepts](https://peleke.github.io/buildlog-template/getting-started/concepts/) | The problem, the claim, and the metric |
+| [CLI Reference](https://peleke.github.io/buildlog-template/guides/cli-reference/) | Every command documented |
+| [MCP Integration](https://peleke.github.io/buildlog-template/guides/mcp-integration/) | Claude Code setup and available tools |
+| [Experiments](https://peleke.github.io/buildlog-template/guides/experiments/) | Running and measuring experiments |
+| [Review Gauntlet](https://peleke.github.io/buildlog-template/guides/review-gauntlet/) | Reviewer personas and the gauntlet loop |
+| [Multi-Agent Setup](https://peleke.github.io/buildlog-template/guides/multi-agent/) | Render rules to any AI coding agent |
+| [Theory](https://peleke.github.io/buildlog-template/theory/00-background/) | The math behind Thompson Sampling |
+| [Philosophy](https://peleke.github.io/buildlog-template/philosophy/) | Principles and honest limitations |
+
+## Contributing
+
+```bash
+git clone https://github.com/Peleke/buildlog-template
+cd buildlog-template
+uv venv && source .venv/bin/activate
+uv pip install -e ".[dev]"
+pytest
+```
+
+We're especially interested in better context representations, credit assignment approaches, statistical methodology improvements, and real-world experiment results (positive or negative).
+
+## License
+
+MIT License — see [LICENSE](./LICENSE)
+
+---
+
+<div align="center">
+
+**"Agent learning" without measurement is just prompt engineering with extra steps.**
+
+**buildlog is measurement.**
+
+</div>

buildlog-0.8.0.dist-info/RECORD

@@ -0,0 +1,54 @@
+buildlog/__init__.py,sha256=wxluyg3fDOiaKUAOJa0cav39hwtNJS5Y_0X20uL_yi4,90
+buildlog/cli.py,sha256=G2d2N2cmC3HEiYc6_mQhRyW8PLHzrHr2Kx38ooQsMlU,59456
+buildlog/confidence.py,sha256=jPvN0_3drGHQG1C7iUxUtYjKC62nNKsHHxu6WXMfJFg,10137
+buildlog/distill.py,sha256=PL7UBToBb27BrCOTWGBTDIXGggtrUumHHBs0_MfG6vY,14166
+buildlog/embeddings.py,sha256=vPydWjJVkYp172zFou-lJ737qsu6vRMQAMs143RGIpA,12364
+buildlog/llm.py,sha256=vXFT03BZVI8GMX-ybSK18SoeABaZbr_B2KXm8WsNews,15278
+buildlog/seeds.py,sha256=L-lzO7rrDjMdVHcYKvEBm6scVDfCmFD0koiyafMZDLo,8596
+buildlog/skills.py,sha256=3CihyyAFfJh0w0VNXQauNhGm_RM8m_GIO1iiNLNPmsQ,33789
+buildlog/stats.py,sha256=2WdHdmzUNGobtWngmm9nA_UmqM7DQeAnZL8_rLQN8aw,13256
+buildlog/core/__init__.py,sha256=gPd6ava_sei0sYIS4NdeHgCVwJ2wEjyLt-Rxz_11IWU,1626
+buildlog/core/bandit.py,sha256=0z58s42jpQiBgXdjb465apenWdBVTvoc7bDOcnAcPLo,24565
+buildlog/core/operations.py,sha256=4gElPJBG9zlXoXO6CPTolwoRrd45DOIV6X1SAQbF3zE,70205
+buildlog/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+buildlog/data/seeds/security_karen.yaml,sha256=Hjwp4j6FB-e0vCXp6AkyhAEkwASR4H1owdhPBID740E,8775
+buildlog/data/seeds/test_terrorist.yaml,sha256=-QY2x6d0LrlM6msMYaYwR43Wny9ZybQmU7pH_E7gSB0,11113
+buildlog/engine/__init__.py,sha256=KNTSUZTUHK6r2abAA-2PIb0wUcRgbE4juvBmoywaXHI,1750
+buildlog/engine/bandit.py,sha256=rfwfQjULakzm-H5bMa5R-iqtI7iibgR4y2JoRhD4zRk,485
+buildlog/engine/confidence.py,sha256=7Zhw8bwT_S-qeVaGXT6dHOyjDTvwyhetLKCWhJH9jMc,656
+buildlog/engine/embeddings.py,sha256=78InITcBbaPZ29tUlb4hJNNVZ1fhxstvVa2fYvOh6QA,597
+buildlog/engine/experiments.py,sha256=38f_UsbY4TP3LTGM9yUsZxoXqxUGOxmLCoYrfZqjVks,20041
+buildlog/engine/types.py,sha256=W7U2B4TE79xACV0cuaidKevmTEskxqXfltH5hjIxV5k,734
+buildlog/mcp/__init__.py,sha256=jCLNUkYFrDcPd5dY9xbaaVANl-ZzdPim1BykgGY7f7U,334
+buildlog/mcp/server.py,sha256=vzgT_-SUWNNx9bvABJBgT7rHNGPx_nB7pVo-04WbwjI,1249
+buildlog/mcp/tools.py,sha256=k1f6R7Dgcx0VIKzTNezkcRc0UVkWXWkk-uGkRq1sIiw,17578
+buildlog/render/__init__.py,sha256=jj-o-vPdqpivdFKK85xskCOT8Zn_2tb3Q5Jb0gWLjAg,2567
+buildlog/render/base.py,sha256=gQfvOsH1zttAo10xtEyNsAbqZ4NRSPiDihO-aiGgTsw,533
+buildlog/render/claude_md.py,sha256=R5z4dHh-3YwclbW3MAzHxgYQHtd5lXS3McJJ1isLGqE,4901
+buildlog/render/continue_dev.py,sha256=S76pGkHssTOEqisiRjKZTfNBRwgiChYeYBR9RKZSgME,3253
+buildlog/render/copilot.py,sha256=8pvXdBWUYnenYIQ5I6L0ukk2JC-NxmNDZrbWQ0kTaqY,3268
+buildlog/render/cursor.py,sha256=tAc0oWJmSXmYzofEzpe1QnkRM-KITTUzHkdSbZXxv4k,3424
+buildlog/render/settings_json.py,sha256=4DS5OWksPrFCa7MIgWIu0t4rxYmItpMdGfTqMX3aMNs,2495
+buildlog/render/skill.py,sha256=_7umIS1Ms1oQ2_PopYueFjX41nMq1p28yJp6DhXFdgU,5981
+buildlog/render/tracking.py,sha256=1y_Mf1s3KblV8TC0W9x2a_ynSJHLwkVZJcW1gs1a6q4,1786
+buildlog/render/windsurf.py,sha256=jfJLb-luQLjdBM38_KKp0mZ08bUSDTBPxSLOIFWxawQ,3047
+buildlog/seed_engine/__init__.py,sha256=f2LYSAs_D1bK4CW_pt54UEmy_SB1u1MOEJlKqM7_R7E,1869
+buildlog/seed_engine/categorizers.py,sha256=uzatJQPG4Vf7z7ZAeyQbL-bVJ8cUKBHG-3U93-BFIRQ,4649
+buildlog/seed_engine/extractors.py,sha256=rkztg62LzFH018ob5Tl1eFp_yfPo8syn3Ib4h6bAyK4,4824
+buildlog/seed_engine/generators.py,sha256=T3N3uHg359xSBypIGmvVWYI0Qh8WLlGIFQWX0_gs0I4,4358
+buildlog/seed_engine/models.py,sha256=qESSuoF3CJy8pp96E3Vmb38V-snzeaW0Xg0RfbH8v1U,3418
+buildlog/seed_engine/pipeline.py,sha256=wLdcjyA_B4wFEb2iKwRCUGkX-oaVu9OXIcmLcEsJN6s,6527
+buildlog/seed_engine/sources.py,sha256=8j9oUFZCSKMr5VpIuAxTPY3wTzfTEmw6M_41_aismiE,11184
+buildlog-0.8.0.data/data/share/buildlog/copier.yml,sha256=hD_UcLGB9eDxTGkWjHAvUoiqQKhLC0w3G-kPKX_p43I,802
+buildlog-0.8.0.data/data/share/buildlog/post_gen.py,sha256=XFlo40LuPpAsBhIRRRtHqvU3_5POss4L401hp35ijhw,1744
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/.gitkeep,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/2026-01-01-example.md,sha256=7x9sKmydfmfKyNz9hV7MtYnQJuBwbxNanbPOcpQDDZQ,7040
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/BUILDLOG_SYSTEM.md,sha256=osclytWwl5jUiTgSpuT4cT3h3oPvCkZ5GPCnFuJZNcY,3802
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/_TEMPLATE.md,sha256=CUvxgcx1-9XT_EdQ8e_vnuPq_h-u1uhXJgForJU2Pso,2932
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/_TEMPLATE_QUICK.md,sha256=eUr5MiqLsM6drV7rAq53R1SLkK8G7LkMAUjWKXx81IA,409
+buildlog-0.8.0.data/data/share/buildlog/template/buildlog/assets/.gitkeep,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+buildlog-0.8.0.dist-info/METADATA,sha256=JNxWkxamVUmEeEQ9vz56j4lzZwex35sJyUngqytW6c4,7032
+buildlog-0.8.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+buildlog-0.8.0.dist-info/entry_points.txt,sha256=BMFclPOomp_sgaa0OqBg6LfqCMlqzjZV88ww5TrPPoo,87
+buildlog-0.8.0.dist-info/licenses/LICENSE,sha256=fAgt-akug9nAwIj6M-SIf8u3ck-T7pJTwfmy9vWYASk,1074
+buildlog-0.8.0.dist-info/RECORD,,