[Docs] Note that /home/ray is not a personal-path leak#63646
Merged
elliot-barn merged 1 commit intoMay 26, 2026
Conversation
Adds a rule to doc/.claude/CLAUDE.md so notebook-output anonymization passes don't treat /home/ray (Ray's runtime home in containers and clusters) as a personal path. Anonymize only real user identifiers and experiment dirs that encode a person or the deprecated AIR runtime. Surfaced during review of ray-project#63464. [DOC-1054] Signed-off-by: Douglas Strodtman <douglas@anyscale.com>
![gemini-code-assist[bot]](https://cdn.statically.io/img/avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Comment on lines
+49
to
+51
| ## Anonymizing paths in notebook output cells | ||
|
|
||
| `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal-path leak — never anonymize it. Anonymize only real user identifiers: personal home prefixes (`/Users/<name>`, `/home/<person>`) become `~`, and experiment or output dirs that encode a person or the deprecated AIR runtime (e.g. `christy-air`) become a neutral equivalent such as `ray_results`. Precedent: #63464. |
Contributor
There was a problem hiding this comment.
The file doc/.claude/CLAUDE.md now exceeds the 50-line limit specified in the header instruction (<!-- Keep under 50 lines. Multi-step procedures → skills. Code style → rules/. -->). Consider condensing this new section to help keep the file size closer to the limit.
Suggested change
| ## Anonymizing paths in notebook output cells | |
| `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal-path leak — never anonymize it. Anonymize only real user identifiers: personal home prefixes (`/Users/<name>`, `/home/<person>`) become `~`, and experiment or output dirs that encode a person or the deprecated AIR runtime (e.g. `christy-air`) become a neutral equivalent such as `ray_results`. Precedent: #63464. | |
| ## Anonymizing paths in notebook output cells | |
| Never anonymize /home/ray (Ray's container/cluster runtime home). Only anonymize real user identifiers: personal home prefixes (/Users/<name>, /home/<person>) to ~, and personal/deprecated AIR experiment dirs (e.g., christy-air) to ray_results. Precedent: #63464. |
elliot-barn
approved these changes
May 26, 2026
dstrodtman
added a commit
to dstrodtman/ray
that referenced
this pull request
May 27, 2026
The anonymized Tune log path in lightgbm_example.ipynb output used /Users/user/ray_results. Switch to ~/ray_results to match the convention in ray-project#63464 and codified in ray-project#63646 (personal home prefixes become ~). [DOC-991] Signed-off-by: Douglas Strodtman <douglas@anyscale.com>
matthewdeng
pushed a commit
that referenced
this pull request
Jun 2, 2026
) ## Update (post-review) Three changes since the original commit, in response to review feedback: 1. **`/home/ray` restored (not a leak).** `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal path. The first commit wrongly anonymized it to `~`; reverted across `batch_tuning`, `pbt_guide`, and `tune-pytorch-lightning` (26 paths). 2. **`christy-air` → `ray_results`.** That experiment dir in `batch_tuning` encodes a person's name plus the deprecated AIR runtime tag, so it stays anonymized — to `ray_results` (Tune's default storage dir), across 18 checkpoint paths. 3. **`pbt_visualization.ipynb` folded in.** Adjacent file with 90 `/Users/rdecal/ray_results` leaks in output cells, anonymized to `~/ray_results`. Brings the total to 10 notebooks. The `/Users/<name>` leaks (kai, rdecal) remain anonymized to `~`. A companion agent rule capturing the `/home/ray` guidance is in #63646. The original description below predates this update; its "substitute `~` for `/home/ray`" method note no longer applies. --- ## Description Cleans up personal-path leaks (`/Users/<name>/...`, `/home/ray/...`) in **output cells** of nine Tune example notebooks under `doc/source/tune/examples/`. 127 leaks removed across 9 files; cell sources untouched. Surfaced by the [DOC-991](https://anyscale1.atlassian.net/browse/DOC-991) (#36167) resolving agent — flagged as adjacent rot during the `pbt_transformers.ipynb` / `lightgbm_example.ipynb` structural fix. ## Related issues [DOC-1054] ## Additional information Method: a one-shot Python script anonymized the leaks (substitute `~` for `/Users/<name>` and `/home/ray` in output-cell text and HTML, preserving per-file JSON indentation 1/2/4-space). Diff is 126±/126± lines across the 9 files, proportional to the original leak count. The 9 affected notebooks: - `ax_example.ipynb` (orthogonal to DOC-1019 ax-platform 1.0.0 API change) - `bayesopt_example.ipynb` (orthogonal to DOC-77 numpy.float deprecation) - `bohb_example.ipynb` - `nevergrad_example.ipynb` - `tune-xgboost.ipynb` - `batch_tuning.ipynb` - `pbt_guide.ipynb` - `tune-pytorch-lightning.ipynb` - `tune_mnist_keras.ipynb` Long-term: leaks will recur until the notebook test/refresh pipeline strips outputs or anonymizes paths before commit. Out of scope for this PR — see DOC-907 for the broader notebook-test-coverage work. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Signed-off-by: Douglas Strodtman <douglas@anyscale.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
rueian
pushed a commit
to rueian/ray
that referenced
this pull request
Jun 4, 2026
…-project#63464) ## Update (post-review) Three changes since the original commit, in response to review feedback: 1. **`/home/ray` restored (not a leak).** `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal path. The first commit wrongly anonymized it to `~`; reverted across `batch_tuning`, `pbt_guide`, and `tune-pytorch-lightning` (26 paths). 2. **`christy-air` → `ray_results`.** That experiment dir in `batch_tuning` encodes a person's name plus the deprecated AIR runtime tag, so it stays anonymized — to `ray_results` (Tune's default storage dir), across 18 checkpoint paths. 3. **`pbt_visualization.ipynb` folded in.** Adjacent file with 90 `/Users/rdecal/ray_results` leaks in output cells, anonymized to `~/ray_results`. Brings the total to 10 notebooks. The `/Users/<name>` leaks (kai, rdecal) remain anonymized to `~`. A companion agent rule capturing the `/home/ray` guidance is in ray-project#63646. The original description below predates this update; its "substitute `~` for `/home/ray`" method note no longer applies. --- ## Description Cleans up personal-path leaks (`/Users/<name>/...`, `/home/ray/...`) in **output cells** of nine Tune example notebooks under `doc/source/tune/examples/`. 127 leaks removed across 9 files; cell sources untouched. Surfaced by the [DOC-991](https://anyscale1.atlassian.net/browse/DOC-991) (ray-project#36167) resolving agent — flagged as adjacent rot during the `pbt_transformers.ipynb` / `lightgbm_example.ipynb` structural fix. ## Related issues [DOC-1054] ## Additional information Method: a one-shot Python script anonymized the leaks (substitute `~` for `/Users/<name>` and `/home/ray` in output-cell text and HTML, preserving per-file JSON indentation 1/2/4-space). Diff is 126±/126± lines across the 9 files, proportional to the original leak count. The 9 affected notebooks: - `ax_example.ipynb` (orthogonal to DOC-1019 ax-platform 1.0.0 API change) - `bayesopt_example.ipynb` (orthogonal to DOC-77 numpy.float deprecation) - `bohb_example.ipynb` - `nevergrad_example.ipynb` - `tune-xgboost.ipynb` - `batch_tuning.ipynb` - `pbt_guide.ipynb` - `tune-pytorch-lightning.ipynb` - `tune_mnist_keras.ipynb` Long-term: leaks will recur until the notebook test/refresh pipeline strips outputs or anonymizes paths before commit. Out of scope for this PR — see DOC-907 for the broader notebook-test-coverage work. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Signed-off-by: Douglas Strodtman <douglas@anyscale.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
limarkdcunha
pushed a commit
to limarkdcunha/ray
that referenced
this pull request
Jun 30, 2026
…3646) ## Description Adds a documentation-authoring rule to `doc/.claude/CLAUDE.md`: `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal-path leak. Notebook-output anonymization passes must not rewrite it. Anonymize only real user identifiers (`/Users/<name>`, `/home/<person>`) and experiment or output dirs that encode a person or the deprecated AIR runtime. ## Related issues Surfaced during review of ray-project#63464, where an anonymization pass over Tune example notebook outputs incorrectly rewrote `/home/ray` to `~`. [DOC-1054] Long-term recurrence prevention (stripping or anonymizing notebook outputs in the test/refresh pipeline) is tracked in DOC-907. ## Additional information The companion content fix — restoring the wrongly-anonymized `/home/ray` paths and anonymizing the leaked `christy-air` experiment dir to `ray_results` — is in ray-project#63464. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Signed-off-by: Douglas Strodtman <douglas@anyscale.com>
limarkdcunha
pushed a commit
to limarkdcunha/ray
that referenced
this pull request
Jun 30, 2026
…-project#63464) ## Update (post-review) Three changes since the original commit, in response to review feedback: 1. **`/home/ray` restored (not a leak).** `/home/ray` is Ray's runtime home directory in containers and clusters, not a personal path. The first commit wrongly anonymized it to `~`; reverted across `batch_tuning`, `pbt_guide`, and `tune-pytorch-lightning` (26 paths). 2. **`christy-air` → `ray_results`.** That experiment dir in `batch_tuning` encodes a person's name plus the deprecated AIR runtime tag, so it stays anonymized — to `ray_results` (Tune's default storage dir), across 18 checkpoint paths. 3. **`pbt_visualization.ipynb` folded in.** Adjacent file with 90 `/Users/rdecal/ray_results` leaks in output cells, anonymized to `~/ray_results`. Brings the total to 10 notebooks. The `/Users/<name>` leaks (kai, rdecal) remain anonymized to `~`. A companion agent rule capturing the `/home/ray` guidance is in ray-project#63646. The original description below predates this update; its "substitute `~` for `/home/ray`" method note no longer applies. --- ## Description Cleans up personal-path leaks (`/Users/<name>/...`, `/home/ray/...`) in **output cells** of nine Tune example notebooks under `doc/source/tune/examples/`. 127 leaks removed across 9 files; cell sources untouched. Surfaced by the [DOC-991](https://anyscale1.atlassian.net/browse/DOC-991) (ray-project#36167) resolving agent — flagged as adjacent rot during the `pbt_transformers.ipynb` / `lightgbm_example.ipynb` structural fix. ## Related issues [DOC-1054] ## Additional information Method: a one-shot Python script anonymized the leaks (substitute `~` for `/Users/<name>` and `/home/ray` in output-cell text and HTML, preserving per-file JSON indentation 1/2/4-space). Diff is 126±/126± lines across the 9 files, proportional to the original leak count. The 9 affected notebooks: - `ax_example.ipynb` (orthogonal to DOC-1019 ax-platform 1.0.0 API change) - `bayesopt_example.ipynb` (orthogonal to DOC-77 numpy.float deprecation) - `bohb_example.ipynb` - `nevergrad_example.ipynb` - `tune-xgboost.ipynb` - `batch_tuning.ipynb` - `pbt_guide.ipynb` - `tune-pytorch-lightning.ipynb` - `tune_mnist_keras.ipynb` Long-term: leaks will recur until the notebook test/refresh pipeline strips outputs or anonymizes paths before commit. Out of scope for this PR — see DOC-907 for the broader notebook-test-coverage work. 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Signed-off-by: Douglas Strodtman <douglas@anyscale.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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.
Description
Adds a documentation-authoring rule to
doc/.claude/CLAUDE.md:/home/rayis Ray's runtime home directory in containers and clusters, not a personal-path leak. Notebook-output anonymization passes must not rewrite it. Anonymize only real user identifiers (/Users/<name>,/home/<person>) and experiment or output dirs that encode a person or the deprecated AIR runtime.Related issues
Surfaced during review of #63464, where an anonymization pass over Tune example notebook outputs incorrectly rewrote
/home/rayto~. [DOC-1054]Long-term recurrence prevention (stripping or anonymizing notebook outputs in the test/refresh pipeline) is tracked in DOC-907.
Additional information
The companion content fix — restoring the wrongly-anonymized
/home/raypaths and anonymizing the leakedchristy-airexperiment dir toray_results— is in #63464.🤖 Generated with Claude Code