Update W&B exporter #21
Merged
Conversation
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
… handling. Updated README for clarity on usage and troubleshooting steps.
Reviewer's GuideThis PR refactors the W&B-to-Neptune migration script by replacing interactive prompts with a full argparse-driven CLI, enhancing logging configuration and error handling (including timeouts and aggregated failure reporting), adding support for uploading console logs and various run-related files to Neptune, migrating metric namespaces and separators to the new Sequence diagram for file upload and logging during run migrationsequenceDiagram
participant Script
participant W&B_Run as W&B Run
participant Neptune_Run as Neptune Run
participant FileSystem as File System
Script->>W&B_Run: List files()
loop For each file
W&B_Run-->>Script: File metadata
Script->>FileSystem: Download file
FileSystem-->>Script: File path
alt output.log
Script->>Neptune_Run: log_string_series(runtime/stdout)
else code/*
Script->>Neptune_Run: assign_files(source_code/files/<filename>)
else requirements.txt
Script->>Neptune_Run: assign_files(source_code/requirements.txt)
else checkpoint/ckpt
Script->>Neptune_Run: assign_files(checkpoints/<filename>)
else other
Script->>Neptune_Run: assign_files(files/<filename>)
end
end
Class diagram for updated file handling and logging in migration scriptclassDiagram
class Run {
+log_configs(dict)
+add_tags(list, group_tags)
+log_string_series(dict, step)
+assign_files(dict)
+log_metrics(dict, step, timestamp)
}
class wandb_run {
+id
+name
+project
+notes
+_attrs
+summary
+scan_history()
+history(stream, pandas)
+files()
+tags
+group
}
class File {
+source
+mime_type
}
class MigrationScript {
+copy_run(wandb_run, wandb_project_name)
+copy_config(neptune_run, wandb_run)
+copy_summary(neptune_run, wandb_run)
+copy_metrics(neptune_run, wandb_run)
+copy_system_metrics(neptune_run, wandb_run)
+copy_console_output(neptune_run, download_path)
+copy_source_code(neptune_run, download_path, filename)
+copy_requirements(neptune_run, download_path)
+copy_other_files(neptune_run, download_path, filename, namespace)
+copy_files(neptune_run, wandb_run, download_folder)
}
MigrationScript --> Run
MigrationScript --> wandb_run
MigrationScript --> File
Class diagram for CLI argument parsing and logging setupclassDiagram
class argparse.ArgumentParser {
+add_argument(...)
+parse_args()
}
class Logger {
+debug(msg)
+info(msg)
+error(msg)
+setLevel(level)
}
class MigrationScript {
+setup_logging(log_level)
}
MigrationScript --> Logger
MigrationScript --> argparse.ArgumentParser
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
…ls on default `num-workers` setting for the W&B to Neptune migration script.
…le-examples into ss/wandb_exporter
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes and found some issues that need to be addressed.
- Avoid relying on a global
tmpdirnameinsidecopy_files—pass the download directory explicitly to keep the function pure and testable. - Using
threadsafe_change_directoryin a multithreaded context can lead to unpredictable cwd changes across threads—prefer uploading files by absolute path instead. - Consider cleaning up the temporary directory after completion (or using a context manager) and exiting with a non-zero code on failures to make automation more robust.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Avoid relying on a global `tmpdirname` inside `copy_files`—pass the download directory explicitly to keep the function pure and testable.
- Using `threadsafe_change_directory` in a multithreaded context can lead to unpredictable cwd changes across threads—prefer uploading files by absolute path instead.
- Consider cleaning up the temporary directory after completion (or using a context manager) and exiting with a non-zero code on failures to make automation more robust.
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:268` </location>
<code_context>
- # filename.replace("code/", ""), wait=True
- # )
- raise NotImplementedError("Source code logging is not implemented in Neptune yet")
+ with threadsafe_change_directory(os.path.join(download_path.replace(filename, ""), "code")):
+ filename = filename.replace("code/", "")
+ neptune_run.assign_files(
</code_context>
<issue_to_address>
Use `os.path.dirname` instead of `replace` for directories
Using string replacement on paths can lead to errors if the filename appears elsewhere in the path. `os.path.dirname(download_path)` is a safer and more reliable way to get the parent directory.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
with threadsafe_change_directory(os.path.join(download_path.replace(filename, ""), "code")):
filename = filename.replace("code/", "")
neptune_run.assign_files(
{f"source_code/files/{filename}": File(source=filename, mime_type="text/plain")}
)
=======
with threadsafe_change_directory(os.path.join(os.path.dirname(download_path), "code")):
filename = filename.replace("code/", "")
neptune_run.assign_files(
{f"source_code/files/{filename}": File(source=filename, mime_type="text/plain")}
)
>>>>>>> REPLACE
</suggested_fix>
### Comment 2
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:282` </location>
<code_context>
- # with threadsafe_change_directory(download_path.replace(filename, "")):
- # neptune_run[namespace].upload_files(filename, wait=True)
- raise NotImplementedError("File logging is not implemented in Neptune yet")
+ with threadsafe_change_directory(download_path.replace(filename, "")):
+ neptune_run.assign_files({f"{namespace}/{filename}": File(source=filename)})
</code_context>
<issue_to_address>
Use `os.path.dirname` to determine the directory
`os.path.dirname(download_path)` is more reliable than using `replace(filename, "")` for extracting the parent directory.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
- with threadsafe_change_directory(download_path.replace(filename, "")):
- neptune_run.assign_files({f"{namespace}/{filename}": File(source=filename)})
=======
+ with threadsafe_change_directory(os.path.dirname(download_path)):
+ neptune_run.assign_files({f"{namespace}/{filename}": File(source=filename)})
>>>>>>> REPLACE
</suggested_fix>
### Comment 3
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:395` </location>
<code_context>
-
- logger, log_filename = setup_logging()
+ wandb_entity = args.wandb_entity
+ neptune_workspace = args.neptune_workspace
+ num_workers = args.num_workers
+ timeout = args.timeout
</code_context>
<issue_to_address>
Validate that Neptune workspace is provided
Add a check to ensure `neptune_workspace` is set, and raise an error if not, to prevent downstream failures.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
wandb_entity = args.wandb_entity
neptune_workspace = args.neptune_workspace
num_workers = args.num_workers
timeout = args.timeout
logger, log_filename = setup_logging(args.log_level)
=======
wandb_entity = args.wandb_entity
neptune_workspace = args.neptune_workspace
num_workers = args.num_workers
timeout = args.timeout
if not neptune_workspace:
raise ValueError("Neptune workspace must be provided via --neptune-workspace to avoid downstream failures.")
logger, log_filename = setup_logging(args.log_level)
>>>>>>> REPLACE
</suggested_fix>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
szaganek
reviewed
Jun 16, 2025
…ved logging and error handling in the migration script to ensure smoother operation and clearer feedback during execution.
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
|
@sourcery-ai review |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes and they look great!
Prompt for AI Agents
Please address the comments from this code review:
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:148` </location>
<code_context>
+ neptune_run.log_configs({f"wandb/{attr}": getattr(wandb_run, attr)})
+ except TypeError:
+ pass
+ except Exception as e:
+ logger.error(
+ f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}' due to exception:\n{e}"
</code_context>
<issue_to_address>
Use `logger.exception` for full traceback
`logger.exception` will log the stack trace, making it easier to debug unexpected errors.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
except Exception as e:
logger.error(
f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}' due to exception:\n{e}"
)
=======
except Exception as e:
logger.exception(
f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}' due to exception:\n{e}"
)
>>>>>>> REPLACE
</suggested_fix>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes - here's some feedback:
- Rather than scattering module‐level globals like
neptune_workspace,timeout, andnum_workers, consider bundling them into a single config or argument object and passing it explicitly to functions for better clarity and testability. - Instead of creating a timestamped directory in the current working directory, you could use Python’s tempfile.TemporaryDirectory to avoid clutter and ensure automatic cleanup.
- Accessing the protected attribute
neptune_run._projectto build the run URL is brittle—switch to a public API or cache the project name separately to prevent breakage if internals change.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Rather than scattering module‐level globals like `neptune_workspace`, `timeout`, and `num_workers`, consider bundling them into a single config or argument object and passing it explicitly to functions for better clarity and testability.
- Instead of creating a timestamped directory in the current working directory, you could use Python’s tempfile.TemporaryDirectory to avoid clutter and ensure automatic cleanup.
- Accessing the protected attribute `neptune_run._project` to build the run URL is brittle—switch to a public API or cache the project name separately to prevent breakage if internals change.
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:347` </location>
<code_context>
try:
- future.result()
+ future.result(timeout=timeout)
+ except TimeoutError:
+ logger.error(f"Timeout copying run {wandb_run.project}/{wandb_run.name}")
+ FAILED_RUNS.append((wandb_run.project, wandb_run.name, "Timeout"))
</code_context>
<issue_to_address>
Catch the correct TimeoutError for futures
Use concurrent.futures.TimeoutError instead of the built-in TimeoutError to correctly catch timeouts from future.result(timeout=...).
</issue_to_address>
### Comment 2
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:254` </location>
<code_context>
- # for line in f:
- # neptune_run["monitoring/stdout"].append(line)
- raise NotImplementedError("Console logging is not implemented in Neptune yet")
+ with open(download_path) as f:
+ for step, line in enumerate(f):
+ neptune_run.log_string_series({"runtime/stdout": line}, step=step)
</code_context>
<issue_to_address>
Explicitly declare encoding when opening files
Not specifying an encoding may cause errors with non-UTF8 files. Use open(download_path, 'r', encoding='utf-8', errors='ignore') or another suitable encoding.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
with open(download_path) as f:
for step, line in enumerate(f):
neptune_run.log_string_series({"runtime/stdout": line}, step=step)
=======
with open(download_path, 'r', encoding='utf-8', errors='ignore') as f:
for step, line in enumerate(f):
neptune_run.log_string_series({"runtime/stdout": line}, step=step)
>>>>>>> REPLACE
</suggested_fix>
### Comment 3
<location> `utils/migration_tools/from_wandb/README.md:58` </location>
<code_context>
+
+### Note
+- If they don't exist yet, Neptune projects corresponding to the W&B projects will be created as [*private*][docs-project-access] with project description set to *Exported from <W&B project URL>*. To change the project visibility or description, use the web app or update the `create_project()` function in `copy_project()`.
+- A `tmp_<timestamp>` directory is created in the current working directory to store the files and logs. Ensure that the current working directory is writable, and that there is enough disk space. You can delete this directory after the script has finished running and the required sanity checks have been performed.
## Metadata mapping from W&B to Neptune
</code_context>
<issue_to_address>
Clarify log file location.
Please specify whether `wandb_to_neptune_<timestamp>.log` is located inside the `tmp_<timestamp>` directory or elsewhere.
</issue_to_address>
### Comment 4
<location> `utils/migration_tools/from_wandb/README.md:57` </location>
<code_context>
+```
+
+### Note
+- If they don't exist yet, Neptune projects corresponding to the W&B projects will be created as [*private*][docs-project-access] with project description set to *Exported from <W&B project URL>*. To change the project visibility or description, use the web app or update the `create_project()` function in `copy_project()`.
+- A `tmp_<timestamp>` directory is created in the current working directory to store the files and logs. Ensure that the current working directory is writable, and that there is enough disk space. You can delete this directory after the script has finished running and the required sanity checks have been performed.
</code_context>
<issue_to_address>
Consider rephrasing guidance on modifying script functions.
Rephrase to clarify that modifying `create_project()` is an advanced customization, not the standard approach, to avoid confusion for typical users.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
- If they don't exist yet, Neptune projects corresponding to the W&B projects will be created as [*private*][docs-project-access] with project description set to *Exported from <W&B project URL>*. To change the project visibility or description, use the web app or update the `create_project()` function in `copy_project()`.
=======
- If they don't exist yet, Neptune projects corresponding to the W&B projects will be created as [*private*][docs-project-access] with project description set to *Exported from <W&B project URL>*. To change the project visibility or description, use the web app. Advanced users may also customize the `create_project()` function in `copy_project()`, but this is not required for typical usage.
>>>>>>> REPLACE
</suggested_fix>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
…irectory usage. Enhance `copy_console_output` function in migration script to log console output to Neptune runs.
|
@sourcery-ai review |
There was a problem hiding this comment.
Pull Request Overview
The PR modernizes the W&B → Neptune migration script by adding full file and console log support, converting interactive prompts to CLI arguments, and standardizing metric namespaces and logging.
- Implements downloading and uploading of run files (logs, source code, requirements, checkpoints, etc.).
- Switches configuration entirely to CLI flags with default values and enhanced logging.
- Renames system metric namespace from
system.*toruntime/*and updates matching documentation.
Reviewed Changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| utils/migration_tools/from_wandb/wandb_to_neptune.py | Add CLI parsing, file and console log copying, update metric namespace and logging |
| utils/migration_tools/from_wandb/README.md | Document new features, changelog, usage instructions, and CLI arguments |
Comments suppressed due to low confidence (5)
utils/migration_tools/from_wandb/wandb_to_neptune.py:331
- Consider renaming the local list
FAILED_RUNStofailed_runsto follow standard snake_case naming for mutable variables.
FAILED_RUNS = []
utils/migration_tools/from_wandb/wandb_to_neptune.py:438
- This local variable
FAILED_PROJECTSshould be renamed tofailed_projectsto adhere to snake_case conventions.
FAILED_PROJECTS = []
utils/migration_tools/from_wandb/wandb_to_neptune.py:164
- Accessing the private attribute
_projectmay break if the internal API changes. Consider using a public property or method onneptune_runto construct the target URL.
f"Copied {wandb_run.url} to https://scale.neptune.ai/{neptune_run._project}/runs/table?viewId=standard-view&query=(%60sys%2Fname%60%3Astring%20MATCHES%20%22{wandb_run.name}%22)&experimentsOnly=false&runsLineage=FULL&lbViewUnpacked=true"
utils/migration_tools/from_wandb/wandb_to_neptune.py:254
- New file-reading and line-by-line logging logic has been introduced here. It would help maintain reliability if unit or integration tests covered
copy_console_outputhandling of missing or malformed log files.
with open(download_path, encoding="utf-8", errors="ignore") as f:
utils/migration_tools/from_wandb/wandb_to_neptune.py:58
- [nitpick] Consider adding a brief docstring explaining the expected format of
log_levelvalues and the return tuple contents (logger,log_filename).
def setup_logging(log_level: str) -> tuple[logging.Logger, str]:
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes - here's some feedback:
- Consider using a tempfile.TemporaryDirectory (or cleaning up
tmpdirnamein a finally block) so that the temporary folder is removed after migration instead of littering the working directory. - Rather than relying on globals like
neptune_workspace,wandb_entity, andtimeoutinsidecopy_project, pass them explicitly as function parameters to improve clarity and testability. - The EXCLUDED_PATHS set mixes prefixes and exact filenames (e.g.
config.yaml); to avoid accidental partial matches you might normalize paths and check equality for files while keeping prefix checks for directories.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Consider using a tempfile.TemporaryDirectory (or cleaning up `tmpdirname` in a finally block) so that the temporary folder is removed after migration instead of littering the working directory.
- Rather than relying on globals like `neptune_workspace`, `wandb_entity`, and `timeout` inside `copy_project`, pass them explicitly as function parameters to improve clarity and testability.
- The EXCLUDED_PATHS set mixes prefixes and exact filenames (e.g. `config.yaml`); to avoid accidental partial matches you might normalize paths and check equality for files while keeping prefix checks for directories.
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:425` </location>
<code_context>
-
- if selected_projects == "":
- selected_projects = wandb_project_names
+ if args.projects:
+ selected_projects = [p.strip() for p in args.projects.split(",") if p.strip()]
else:
- selected_projects = [project.strip() for project in selected_projects.split(",")]
+ selected_projects = wandb_project_names
if not_found := set(selected_projects) - set(wandb_project_names):
</code_context>
<issue_to_address>
No check for empty selected_projects after filtering.
Consider exiting or prompting the user if selected_projects is empty to prevent unnecessary execution.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
if args.projects:
selected_projects = [p.strip() for p in args.projects.split(",") if p.strip()]
else:
selected_projects = wandb_project_names
if not_found := set(selected_projects) - set(wandb_project_names):
print(f"Projects not found: {not_found}")
print(f"Copying {len(selected_projects)} projects: {selected_projects}\n")
logger.info(f"Copying {len(selected_projects)} projects: {selected_projects}")
=======
if args.projects:
selected_projects = [p.strip() for p in args.projects.split(",") if p.strip()]
else:
selected_projects = wandb_project_names
if not selected_projects:
print("No valid projects selected after filtering. Exiting.")
exit(1)
if not_found := set(selected_projects) - set(wandb_project_names):
print(f"Projects not found: {not_found}")
print(f"Copying {len(selected_projects)} projects: {selected_projects}\n")
logger.info(f"Copying {len(selected_projects)} projects: {selected_projects}")
>>>>>>> REPLACE
</suggested_fix>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes - here's some feedback:
- Refactor
copy_runinto smaller subfunctions for attribute handling, metrics copying, and file uploads to reduce complexity and improve readability. - Avoid relying on module-level globals (e.g.,
neptune_workspace,timeout) inside core functions by passing them explicitly as parameters to improve modularity and testability. - Consider batching or streaming console log lines rather than logging each line individually to reduce API calls and improve performance when handling large log files.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Refactor `copy_run` into smaller subfunctions for attribute handling, metrics copying, and file uploads to reduce complexity and improve readability.
- Avoid relying on module-level globals (e.g., `neptune_workspace`, `timeout`) inside core functions by passing them explicitly as parameters to improve modularity and testability.
- Consider batching or streaming console log lines rather than logging each line individually to reduce API calls and improve performance when handling large log files.
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:283` </location>
<code_context>
+
+
+def copy_files(neptune_run: Run, wandb_run, download_folder: str) -> None: # type: ignore
+ for file in wandb_run.files():
+ if file.size and not any(file.name.startswith(path) for path in EXCLUDED_PATHS):
+ download_path = os.path.join(download_folder, file.name)
+ try:
+ logger.debug(
+ f"Downloading file {file.name} for run {wandb_run.name} to {download_path}"
+ )
+ file.download(root=download_folder, replace=True, exist_ok=True)
+ if file.name == "output.log":
</code_context>
<issue_to_address>
No handling for file name collisions in copy_files.
Files with identical names from different subdirectories may overwrite each other in the download_folder. Please implement logic to preserve directory structure or handle name collisions to prevent data loss.
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
There was a problem hiding this comment.
Hey @SiddhantSadangi - I've reviewed your changes - here's some feedback:
- Remove unused helper functions and commented-out code (e.g., the threadsafe_change_directory context manager) to reduce clutter and maintainability burden.
- Add cleanup logic to delete the temporary directory after migration completes to prevent leftover files from accumulating on disk.
- Rename local mutable lists FAILED_RUNS and FAILED_PROJECTS to lowercase (e.g., failed_runs, failed_projects) to follow Python naming conventions for non-constant variables.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Remove unused helper functions and commented-out code (e.g., the threadsafe_change_directory context manager) to reduce clutter and maintainability burden.
- Add cleanup logic to delete the temporary directory after migration completes to prevent leftover files from accumulating on disk.
- Rename local mutable lists FAILED_RUNS and FAILED_PROJECTS to lowercase (e.g., failed_runs, failed_projects) to follow Python naming conventions for non-constant variables.
## Individual Comments
### Comment 1
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:347` </location>
<code_context>
try:
- future.result()
+ future.result(timeout=timeout)
+ except TimeoutError:
+ logger.error(f"Timeout copying run {wandb_run.project}/{wandb_run.name}")
+ FAILED_RUNS.append((wandb_run.project, wandb_run.name, "Timeout"))
</code_context>
<issue_to_address>
Ambiguous TimeoutError catch
`future.result(timeout=timeout)` raises `concurrent.futures.TimeoutError`, not the built-in `TimeoutError`. Please import and catch `concurrent.futures.TimeoutError` explicitly.
</issue_to_address>
### Comment 2
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:46` </location>
<code_context>
+EXCLUDED_PATHS = {"artifact/", "config.yaml", "media/", "wandb-"}
+
def exc_handler(exctype, value, tb):
logger.exception("".join(traceback.format_exception(exctype, value, tb)))
</code_context>
<issue_to_address>
exc_handler never registered
Set `sys.excepthook = exc_handler` after defining the handler to ensure it processes uncaught exceptions.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
def exc_handler(exctype, value, tb):
logger.exception("".join(traceback.format_exception(exctype, value, tb)))
=======
def exc_handler(exctype, value, tb):
logger.exception("".join(traceback.format_exception(exctype, value, tb)))
import sys
sys.excepthook = exc_handler
>>>>>>> REPLACE
</suggested_fix>
### Comment 3
<location> `utils/migration_tools/from_wandb/wandb_to_neptune.py:149` </location>
<code_context>
+ except TypeError:
+ pass
+ except Exception as e:
+ logger.error(
+ f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}' due to exception:\n{e}"
+ )
</code_context>
<issue_to_address>
Use logger.exception to capture traceback
Using `logger.exception` will automatically include the stack trace, making debugging easier.
</issue_to_address>
<suggested_fix>
<<<<<<< SEARCH
except Exception as e:
logger.error(
f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}' due to exception:\n{e}"
)
=======
except Exception:
logger.exception(
f"[{copy_run.__name__}]\tFailed to copy '{attr}' from W&B run '{wandb_run.name}'"
)
>>>>>>> REPLACE
</suggested_fix>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
szaganek
previously approved these changes
Jun 16, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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
runtimefromsystem..in system metric names with/to match the Neptune metric namespace format.Any expected test failures?
Add a
[X]to relevant checklist items❔ This change
✔️ Pre-merge checklist
🧪 Test Configuration
Summary by Sourcery
Enhance the W&B to Neptune migration script by adding file and console log support, converting interactive prompts to CLI arguments, refining metric namespaces, and improving logging, error handling, and configuration options.
New Features:
Enhancements:
systemtoruntimeand replace dots with slashes in metric names--num-workersto 1 and expose--log-levelfor configurable logging verbosityDocumentation: