fix: counts api response structure change

[nikhilsinhaparseable]

Summary by CodeRabbit

• Bug Fixes
  • Improved timestamp formatting in count responses with timezone information, ensuring consistency across all queries.
  • Enhanced handling of grouped count queries with proper column value mapping and data normalization.
  • Standardized response field names from camelCase to snake_case for improved API consistency and predictability.

[coderabbitai]

Walkthrough

The get_counts handler is refactored to handle group-by columns in filtered queries. It normalizes timestamp formatting in SQL results, reconstructs Arrow record batches with standardized timestamps, and enriches count records with corresponding group-by column values. The non-filtered code path also updates field naming for consistency.

Changes

Cohort / File(s)	Summary
Count Handler Query Logic src/handlers/http/query.rs	Modified get_counts to detect group-by columns, normalize timestamps by appending +00:00 to start_time/end_time fields via CountsRecord conversion, reconstruct Arrow RecordBatch with normalized values, and enrich results with group-by column values from original JSON records. Also updated non-filtered path field names from endTime to end_time.

Sequence Diagram

sequenceDiagram
    participant Client
    participant Handler as HTTP Handler
    participant SQL as SQL Engine
    participant Convert as JSON/Record Converter
    participant Process as Record Post-Processor
    
    Client->>Handler: GET counts (with group_by, conditions)
    Handler->>SQL: Execute counts query (includes group_by columns)
    SQL-->>Handler: Return RecordBatch with results
    
    Handler->>Convert: Convert RecordBatch to JSON
    Convert-->>Handler: JSON records
    
    Handler->>Process: Parse records to CountsRecord objects
    Process->>Process: Normalize timestamps (+00:00)
    Process-->>Handler: Normalized CountsRecord objects
    
    Handler->>Convert: Construct new RecordBatch<br/>(start_time, end_time, count)
    Convert-->>Handler: Normalized RecordBatch
    
    Handler->>Convert: Convert back to JSON
    Convert-->>Handler: Normalized JSON records
    
    alt group_by columns present
        Handler->>Process: Enrich records with group-by values<br/>(copy from original JSON)
        Process-->>Handler: Enriched count records
    end
    
    Handler-->>Client: Return counts response

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• #1601: Adds selection and grouping of group_by columns in counts SQL queries, which this PR consumes by enriching results with group-by column values.

Suggested reviewers

• parmesant

Poem

🐰 Hop, hop, timestamps align!
Group-by treasures now entwine,
Arrow records dance and glow,
Enriched counts steal the show! ✨

🚥 Pre-merge checks | ✅ 1 | ❌ 2

❌ Failed checks (2 warnings)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	No pull request description was provided by the author, missing required sections including Description, rationale, key changes, and testing/documentation checkboxes.	Add a comprehensive description explaining what the counts API response structure change fixes, why it was needed, how it impacts consumers, and confirm the required testing and documentation checklist items.
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: fixing the counts API response structure, which aligns with the modifications to response field names and record structure.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

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

Inline comments:
In `@src/handlers/http/query.rs`:
- Around line 390-396: The code currently allows group_by columns (group_by_cols
/ body.conditions.group_by) to include reserved synthesized keys ("start_time",
"end_time", "count", "end_time") which later causes rec.insert(...) to clobber
normalized count payload and duplicate keys in fields; add validation after
constructing group_by_cols (before using has_groups/rec.insert) to detect any
intersection with the reserved set {"start_time","end_time","count"} and return
a clear BadRequest/validation error rejecting the request (or alternatively
alias those names) so collisions cannot reach rec.insert; reference
group_by_cols, body.conditions.group_by, has_groups, rec.insert and fields when
implementing the check and error path.
- Around line 424-446: The serde_json::from_value(...) and
RecordBatch::try_from_iter(...) error paths should be mapped to a 5xx internal
error instead of client errors; change the serde_json::from_value(...) call
(producing CountsRecord) to map_err into the QueryError variant used for
server/internal failures (e.g., QueryError::InternalError) and likewise change
the .map_err on RecordBatch::try_from_iter(...) to the same 5xx-mapped
QueryError variant, preserving the original error text in the message; update
the two call sites (serde_json::from_value and RecordBatch::try_from_iter) to
use the internal-error QueryError variant rather than QueryError::SerdeJsonError
or QueryError::CustomError so server-side transform failures return 5xx.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: f04d7a6a-a3c4-4dd0-b041-eebffc40795a

📥 Commits

Reviewing files that changed from the base of the PR and between 79656f6 and 417f1de.

📒 Files selected for processing (1)

• src/handlers/http/query.rs