Track Peer Repos

Track recent activity in normalizing flow and density estimation open-source repositories.

Arguments: $ARGUMENTS

Tracked Repositories

The following are key repositories in the normalizing flow ecosystem:

Short Name	GitHub Repo	Description
nflows	bayesiains/nflows	Normalizing flows in PyTorch
normflows	VincentStimper/normalizing-flows	Normalizing flows library
FrEIA	vislearn/FrEIA	Framework for Easily Invertible Architectures
BNAF	nicola-decao/BNAF	Block Neural Autoregressive Flow
survae-flows	didriknielsen/survae_flows	Surjective and surjective flows
zuko	probabilists/zuko	Normalizing flows in PyTorch
flowtorch	facebookresearch/flowtorch	Flow-based generative models
neural-spline-flows	bayesiains/nsf	Neural Spline Flows
FFJORD	rtqichen/ffjord	Free-form Jacobian of Reversible Dynamics
residual-flows	rtqichen/residual-flows	Residual Flows

Task: Generate Activity Report

1. Parse Arguments

Parse $ARGUMENTS to extract:

•Time range: e.g., "7 days ago", "1 month ago", "2025-01-01". Default to "14 days ago" if not specified.
•Repo filter: If specific repo names are given, only track those. Otherwise track all.

2. Collect Data for Each Repository

For each repository, use gh CLI to gather:

bash

# Recent commits
gh api repos/{owner}/{repo}/commits --jq '.[0:20] | .[] | "\(.sha[0:7])|\(.commit.author.date)|\(.commit.message | split("\n")[0])"' 2>/dev/null

# Recent releases
gh api repos/{owner}/{repo}/releases --jq '.[0:5] | .[] | "\(.tag_name)|\(.published_at)|\(.name)"' 2>/dev/null

# Open issues count and recent issues
gh api repos/{owner}/{repo}/issues?state=open\&per_page=5 --jq '.[] | "\(.number)|\(.created_at)|\(.title)"' 2>/dev/null

# Repository stats
gh api repos/{owner}/{repo} --jq '"\(.stargazers_count) stars | \(.forks_count) forks | \(.open_issues_count) open issues"' 2>/dev/null

3. Analyze Activity

For each repository, assess:

Activity Level

•Commit frequency: commits per week in the time range
•Release cadence: any new releases?
•Issue activity: new issues opened/closed?
•Rating: Dormant / Low / Moderate / Active / Very Active

Technical Direction

Categorize commits by type:

•feat: New features, capabilities, model architectures
•fix: Bug fixes, numerical stability improvements
•perf: Performance optimizations, memory efficiency
•refactor: Code restructuring, API changes
•docs: Documentation updates
•test: Test additions/improvements

Relevance to BreezeForest

For significant updates, note:

•Does it relate to autoregressive flows?
•Any improvements to Jacobian computation?
•New training techniques or regularization methods?
•Sampling/inversion improvements?

4. Generate Report

Output the report in this format:

markdown

# Peer Repository Activity Report
**Period**: [start date] to [end date]
**Generated**: YYYY-MM-DD

## Activity Overview

| Repository | Stars | Commits | Releases | Activity |
|---|---|---|---|---|
| nflows | ⭐ XXX | N commits | vX.Y.Z | 🟢 Active |
| normflows | ⭐ XXX | N commits | - | 🟡 Moderate |
| ... | ... | ... | ... | ... |

## Detailed Analysis

### [Repository Name]
**Activity**: [rating]
**Stats**: [stars] stars, [forks] forks

#### Key Updates
- [commit hash] [description] — **[category]**: [brief analysis]
- ...

#### Technical Direction
[Summary of where this project is heading]

#### Relevance to BreezeForest
[What BreezeForest can learn from or compare against]

### [Next repository...]

## Cross-Project Trends
1. [Trend observed across multiple repos]
2. ...

## Signals Worth Noting

### Notable
- [Signal]: [description + implication]

### Opportunities
- [Opportunity]: [description + how BreezeForest could benefit]

## Metadata
- Report generated: YYYY-MM-DD
- Data range: [start] to [end]
- Repositories tracked: [count]

5. Save Report

Save to notes/peer_repos/activity_YYYY_MM_DD.md. Create the directory if it doesn't exist.