Sxnan opened a new issue, #777:
URL: https://github.com/apache/flink-agents/issues/777

   ### Search before asking
   
   - [x] I searched in the 
[issues](https://github.com/apache/flink-agents/issues) and found nothing 
similar.
   
   ### Description
   
   Found during integrate_with_flink doc verification (#743).
   
   Collection of doc-quality improvements for
   `docs/content/docs/development/integrate_with_flink.md`:
   
   1. **Overview grammar** (line 25):
      > By integrate agent with flink DataStream/Table, Flink Agents can 
leverage the powerful data processing ability of Flink.
   
      Should read "By integrating agent with Flink DataStream/Table". `Flink`
      needs capitalization too.
   
   2. **Snippets are not self-contained** — every Python and Java snippet uses
      undefined symbols and ellipses, so a new user cannot copy-and-run them:
      - `YourPojo`, `MyKeySelector` introduced without definition (lines 78, 
83, 144)
      - `env.from_source(...)` / `env.fromSource(...)` / 
`t_env.from_elements(...)`
        left as `...`
      - Imports missing for `ExternalTypeInfo`, `RowTypeInfo`, `BasicTypeInfo`,
        `Schema`, `DataTypes`, `KeySelector`
   
      Either show runnable forms (e.g. `env.from_collection([...])` and a
      concrete `MyKeySelector` class) or add a link to the matching example
      under `examples/...` and `python/flink_agents/examples/...`.
   
   3. **Java Table example: misleading "always nested row" comment** (lines 
158-159):
      ```java
      // Here the output schema should always be a nested row, of which
      // the f0 column is the expected row.
      Schema outputSchema =
              Schema.newBuilder()
                      .column("f0", DataTypes.ROW(DataTypes.FIELD("result", 
DataTypes.DOUBLE())))
                      .build();
      ```
      The integration test `FlinkIntegrationTest.testFromTableToTable:140` uses
      a flat `Schema.newBuilder().column("f0", DataTypes.STRING())` for a String
      output. "Always" is wrong — whether the row needs to be nested depends on
      what the agent emits per record. Rephrase to explain when nesting is
      required.
   
   4. **Python vs Java column naming inconsistency not explained** — the Python
      `to_table` example (line 141) uses `column("result", DataTypes.INT())` and
      a matching `RowTypeInfo([...], ["result"])`, while the Java tab forces
      `f0`. This is a real implementation difference (Python's
      `RemoteAgentBuilder.to_table` re-routes through `PickledBytesTypeInfo`;
      Java's `toTable(Schema)` calls 
`tableEnv.fromDataStream(DataStream<Object>, schema)`
      which yields the `f0` default column name). Either harmonize the runtime
      behavior or document why the schemas look different.
   
   ### Where to fix
   
   `docs/content/docs/development/integrate_with_flink.md`. For (4) a follow-up
   runtime-behavior issue may be appropriate if the team wants to converge the
   two paths instead of just documenting the asymmetry.
   
   ### Version and environment
   
   Flink Agents 0.3.0 (`main`).
   
   ### Are you willing to submit a PR?
   
   - [ ] I'm willing to submit a PR!
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to