Update AI infrastructure: CLAUDE.md, .gitignore#2938
Open
jamesblackburn wants to merge 4 commits intomasterfrom
Open
Update AI infrastructure: CLAUDE.md, .gitignore#2938jamesblackburn wants to merge 4 commits intomasterfrom
jamesblackburn wants to merge 4 commits intomasterfrom
Conversation
jamesblackburn
requested review from
IvoDD,
alexowens90,
maxim-morozov and
poodlewars
jamesblackburn
commented
Feb 25, 2026
docs/claude/skills/code-review.md
Outdated
Collaborator
Author
There was a problem hiding this comment.
Benefits:
- A bunch of ArcticDB based edge-case handling for claude to worry / think about.
- Runs the different analyses in parallel in sub-agents
jamesblackburn
commented
Feb 25, 2026
| ArcticDB is a high-performance, serverless DataFrame database for the Python Data Science ecosystem. It provides a Python API backed by a C++ data-processing and compression engine, supporting S3, LMDB, Azure Blob Storage, and MongoDB backends. | ||
|
|
||
| ## Claude-Maintained Documentation | ||
| ## Documentation |
Collaborator
Author
There was a problem hiding this comment.
Brought the doc updates section together: dev docs and user-docs.
jamesblackburn
added
documentation
jamesblackburn
force-pushed
the
jb/2_infra_claude_docs
branch
2 times, most recently
from
b9ca1b2 to
30a0f61
Compare
jamesblackburn
commented
Feb 25, 2026
| ARCTICDB_PROTOC_VERS=4 CMAKE_BUILD_PARALLEL_LEVEL=16 ARCTIC_CMAKE_PRESET=linux-debug pip install -ve . | ||
| ``` | ||
|
|
||
| To install packages which aren't available internally, use the following custom index: |
Collaborator
Author
There was a problem hiding this comment.
How to resolve specific package versions
jamesblackburn
commented
Feb 25, 2026
|
|
||
| ## Benchmarking | ||
|
|
||
| **IMPORTANT: Always use a release build for benchmarking.** Debug builds have 10-30x overhead from disabled optimizations, assertions, and unoptimized template instantiation (e.g. sparrow/Arrow type system). Use `ARCTIC_CMAKE_PRESET=linux-release` for both C++ and Python benchmarks. |
Collaborator
Author
There was a problem hiding this comment.
This is important or it sometimes runs benchmarks with a debug build
jamesblackburn
commented
Feb 25, 2026
jamesblackburn
force-pushed
the
jb/2_infra_claude_docs
branch
from
30a0f61 to
1cde8c5
Compare
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
Contributor
ArcticDB Code Review SummaryThis PR makes documentation and infrastructure changes to CLAUDE.md, .gitignore, PR_REVIEW_GUIDELINES.md, CpuCount.cmake, setup.py, and .cproject. No public Python API, storage backend, or on-disk format changes. What Changed (all commits)
Delta Analysis (6e65a7d - merge of master)No PR-owned code changed. The merge incorporates four master commits with no merge conflicts. Checklist items are unaffected by this delta. Previously Flagged Issues
ChecklistAPI & Compatibility
Memory & Safety
Correctness
Code Quality
Testing
Build & Dependencies
Security
PR Title & Description
Documentation
|
jamesblackburn
changed the title
Collaborator
maxim-morozov
left a comment
maxim-morozov
left a comment
There was a problem hiding this comment.
Couple of comments, but LGTM
jamesblackburn
force-pushed
the
jb/2_infra_claude_docs
branch
from
0378f3f to
d86f4aa
Compare
jamesblackburn
commented
Mar 2, 2026
|
|
||
| --- | ||
|
|
||
| ## 21. DOCUMENTATION |
Collaborator
Author
There was a problem hiding this comment.
Added documentation review for the code-review skill
jamesblackburn
commented
Mar 2, 2026
| if(${_needed_mb} GREATER ${_physical_mb}) | ||
| message(WARNING "We recommend 1500MB of physical RAM per processor core. We found ${_proc_count} cores and\ | ||
| ${_physical_mb}MB of RAM. Use the CMAKE_BUILD_PARALLEL_LEVEL environment variable to avoid thrashing.") | ||
| # C++ compiler processes can use up to 3GB each; cap parallelism to avoid OOM |
Collaborator
Author
There was a problem hiding this comment.
Fix periodic OOM on large-core client machine. Consider machine memory for parallel builds by default. cpp compiler commands can use op to 3GB
jamesblackburn
force-pushed
the
jb/2_infra_claude_docs
branch
from
5463ba0 to
8d12ad7
Compare
- Add documentation guidance: NumPy docstring format, tutorial and mkdocs.yml checklist for new features - Expand ASV benchmarking section with suite names, env vars, and run-from-root/--python caveat; simplify by removing general ASV knowledge Claude already has - Remove CMAKE_BUILD_PARALLEL_LEVEL from build commands (now automatic) - Add TDD guidelines, git workflow (confirm before push), branch work log convention, and code formatter scoping guidance - Broaden .ipynb_checkpoints/ gitignore pattern to any directory level
Replace the RAM warning with actual job limiting: caps parallelism at RAM_MB/3000 (~3GB per compiler process) to prevent OOM. Uses message(NOTICE) for the informational message so it goes to stderr and doesn't break the CI step that pipes cmake stdout to \$GITHUB_ENV. Updates setup.py to parse the cmake STATUS message prefix correctly. Removes redundant .cproject source path entry.
Add §21 DOCUMENTATION covering NumPy-format docstring completeness, tutorial requirements for complex features, mkdocs.yml nav updates, edge case/limitation coverage, and docs/claude/ update obligations. Expand the checklist from 2 items to 6 to align with CLAUDE.md.
jamesblackburn
force-pushed
the
jb/2_infra_claude_docs
branch
from
8d12ad7 to
8f4e8d2
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Update AI infrastructure: CLAUDE.md, build system, and review skill
Improves developer/AI workflow documentation, fixes build parallelism, and aligns the code
review skill with documentation requirements.
Changes
CLAUDE.mdchecklist)
available suites,
--python=$(which python)usageCMAKE_BUILD_PARALLEL_LEVEL=16from wheel build command (now setautomatically)
cpp/CMake/CpuCount.cmakeRAM_MB / 3000(~3 GB per compiler process) to prevent OOM during buildssetup.py-- CMAKE_BUILD_PARALLEL_LEVEL=prefix emitted byCMake STATUS messages
docs/claude/PR_REVIEW_GUIDELINES.mdfor complex features,
mkdocs.ymlnav updates, edge case/limitation coverage, anddocs/claude/update obligations when modifying documented areasCLAUDE.md.gitignore.ipynb_checkpoints/pattern to match at any directory level.cproject<pathentry kind="src" path=""/>line