src/sampleworks/utils/guidance_script_utils.py (1)

584-584: Remove the redundant inline comment.

Line 584 restates what the code already makes clear; dropping it keeps this file closer to the repo’s “direct, readable” style.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/sampleworks/utils/guidance_script_utils.py` at line 584, Remove the
redundant inline comment that repeats the code’s action ("# write out the job
parameters to a JSON file in the same directory as the refined.cif file");
delete that comment line near the code that writes job parameters to JSON (the
block referencing refined.cif/job parameters) so the implementation (in
guidance_script_utils.py) remains direct and uncluttered.

src/sampleworks/utils/guidance_script_utils.py (1)

584-587: ⚠️ Potential issue | 🟠 Major

Previously flagged issues remain unresolved; additionally, consider including execution metadata.

The directory creation and serialization concerns from the previous review are still valid and unaddressed:

1. output_dir may not exist if _run_guidance() fails before save_everything() runs
2. json.dump(job.__dict__, ...) will raise TypeError for non-serializable fields (e.g., GuidanceType, StructurePredictor objects)
3. Any exception aborts the entire job queue

New observation: Writing only job.__dict__ omits valuable execution results. Per the JobResult dataclass, fields like status, exit_code, runtime_seconds, started_at, and finished_at are only in job_result, not in job. Consider merging both for complete metadata.

Proposed fix incorporating both concerns

-        # write out the job parameters to a JSON file in the same directory as the refined.cif file
-        with open(Path(job_result.output_dir) / "job_metadata.json", "w") as fp:
-            json.dump(job.__dict__, fp)
+        # write out the job parameters and execution results to a JSON file
+        metadata_path = Path(job_result.output_dir) / "job_metadata.json"
+        try:
+            metadata_path.parent.mkdir(parents=True, exist_ok=True)
+            combined_metadata = {
+                "config": job.__dict__,
+                "result": job_result.__dict__,
+            }
+            with metadata_path.open("w", encoding="utf-8") as fp:
+                json.dump(combined_metadata, fp, indent=2, default=str)
+        except (OSError, TypeError) as exc:
+            logger.warning(f"Failed to write job metadata to {metadata_path}: {exc}")

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/sampleworks/utils/guidance_script_utils.py` around lines 584 - 587,
Ensure the job metadata write is robust: before opening
Path(job_result.output_dir) / "job_metadata.json" ensure the directory exists
(create parents with exist_ok=True) and perform the write atomically to avoid
partial files; merge serializable representations of job and job_result (use
job.__dict__ plus JobResult fields like status, exit_code, runtime_seconds,
started_at, finished_at) into a single dict; convert or filter non-serializable
values from GuidanceType/StructurePredictor (e.g., replace with their names/IDs
or call a to_dict() if available) or provide a default JSON serializer to avoid
TypeError when json.dump is called; wrap the file write in a try/except that
logs the exception (reference process/logger used elsewhere) instead of raising
so a single metadata write failure does not abort the entire job queue (locate
code around _run_guidance(), save_everything(), job_result.output_dir, and the
JobResult dataclass to implement these changes).

src/sampleworks/utils/guidance_script_utils.py (1)

587-590: ⚠️ Potential issue | 🟠 Major

Prevent metadata-write I/O failures from aborting the remaining job queue.

If this write hits an OSError (e.g., permission/disk issues), the loop exits and skips remaining jobs. Since metadata is auxiliary, handle write failures and continue.

Proposed fix

-        # write out the job parameters to a JSON file in the same directory as the refined.cif file
-        with open(Path(job_result.output_dir) / "job_metadata.json", "w") as fp:
-            json.dump(job.__dict__, fp)
+        # write out the job parameters to a JSON file in the same directory as the refined.cif file
+        metadata_path = Path(job_result.output_dir) / "job_metadata.json"
+        metadata_path.parent.mkdir(parents=True, exist_ok=True)
+        try:
+            with metadata_path.open("w", encoding="utf-8") as fp:
+                json.dump(job.__dict__, fp)
+        except OSError as exc:
+            logger.warning(f"Failed to write job metadata to {metadata_path}: {exc}")

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/sampleworks/utils/guidance_script_utils.py` around lines 587 - 590, The
metadata write (opening Path(job_result.output_dir) / "job_metadata.json" and
json.dump(job.__dict__, fp)) can raise OSError and currently aborts the job
loop; wrap the open/json.dump in a try/except that catches OSError (or OSError
and IOError for compatibility), log the error with context including
job_result.output_dir and job identifier (e.g., job.id or other distinguishing
field) using the module's logger, and continue without re-raising so remaining
jobs are processed. Ensure the except only swallows I/O-related exceptions and
does not hide other failures.