#855 Add COMP-1 and COMP-2 floating point encoders for the writer.

[yruslan]

Closes #855

Summary by CodeRabbit

• New Features

  • Added configurable floating-point format support for COBOL COMP-1 and COMP-2 encoding, including IEEE 754 and IBM variants with selectable endianness (and consistent handling of special values like NaN/infinities).

• Bug Fixes

  • Ensured the selected floating-point format is correctly applied during primitive encoding.

• Tests

  • Added byte-level encoder validation across IEEE 754 and IBM formats, plus writer tests covering multiple floating_point_format options with edge-case coverage.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 92dbb085-e989-4a4b-901f-0646b0fa8b2b

📥 Commits

Reviewing files that changed from the base of the PR and between 0e4a2d5 and 124bcad.

📒 Files selected for processing (2)

• cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/EncoderSelector.scala
• cobol-parser/src/test/scala/za/co/absa/cobrix/cobol/parser/encoding/FloatingPointEncodersSuite.scala

🚧 Files skipped from review as they are similar to previous changes (1)

• cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/EncoderSelector.scala

---

Walkthrough

Adds write support for COBOL COMP-1 and COMP-2 floating-point fields. A new FloatingPointEncoders object provides byte-array encoders for IEEE-754 and IBM hex-float formats in both endiannesses. EncoderSelector.getEncoder gains a floatingPointFormat parameter and new dispatch cases routing COMP1/COMP2 decimal types to those encoders. ParserVisitor is updated to thread floatingPointFormat through. Comprehensive unit and integration tests cover all four encoding formats.

Changes

COMP-1/COMP-2 floating-point write support

Layer / File(s)	Summary
FloatingPointEncoders: IEEE-754 and IBM hex-float byte encoders cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/FloatingPointEncoders.scala	New singleton object with eight encoding methods: IEEE-754 single/double in both endiannesses via ByteBuffer, and IBM hex-float single/double big-endian with special-case handling for 0.0, NaN, and infinities plus base-16 exponent/fraction packing; little-endian IBM variants produced by byte reversal.
EncoderSelector: COMP1/COMP2 dispatch and ParserVisitor wiring cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/EncoderSelector.scala, cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/antlr/ParserVisitor.scala	getEncoder receives a new floatingPointFormat parameter (default IBM); two new match arms route Decimal(COMP1)/Decimal(COMP2) to getSingleFloatingPointEncoder/getDoubleFloatingPointEncoder helpers that convert numeric inputs and delegate to FloatingPointEncoders. ParserVisitor.visitPrimitive passes floatingPointFormat to getEncoder.
FloatingPointEncodersSuite: unit tests for floating-point encoders cobol-parser/src/test/scala/za/co/absa/cobrix/cobol/parser/encoding/FloatingPointEncodersSuite.scala	Test suite validating all eight encoder functions across IEEE-754 and IBM formats: exact-byte assertions for zero, one, negative one, infinities, NaN, and MaxValue/MinPositiveValue; round-trip decode validation; and byte-position/endianness verification for sign, exponent, and fraction fields.
FixedLengthEbcdicWriterSuite: COMP-1/COMP-2 encoding integration tests spark-cobol/src/test/scala/za/co/absa/cobrix/spark/cobol/writer/FixedLengthEbcdicWriterSuite.scala	Four integration test cases validate each floating_point_format option (IEEE754_little_endian, IEEE754, IBM_little_endian, IBM) by writing DataFrames with positive/negative/zero/infinity/NaN/null float values and asserting written bytes match pre-computed expected arrays.

Sequence Diagram(s)

sequenceDiagram
    participant ParserVisitor
    participant EncoderSelector
    participant SingleFloatHelper as getSingle<br/>FloatingPoint<br/>Encoder
    participant DoubleFloatHelper as getDouble<br/>FloatingPoint<br/>Encoder
    participant FloatingPointEncoders

    ParserVisitor->>EncoderSelector: getEncoder(Decimal(COMP1), floatingPointFormat=IEEE754)
    EncoderSelector->>SingleFloatHelper: convert input to Float
    SingleFloatHelper->>FloatingPointEncoders: encodeIeee754SingleBigEndian(f)
    FloatingPointEncoders-->>SingleFloatHelper: Array[Byte]
    SingleFloatHelper-->>EncoderSelector: Encoder
    EncoderSelector-->>ParserVisitor: Some(Encoder)

    ParserVisitor->>EncoderSelector: getEncoder(Decimal(COMP2), floatingPointFormat=IBM)
    EncoderSelector->>DoubleFloatHelper: convert input to Double
    DoubleFloatHelper->>FloatingPointEncoders: encodeIbmDoubleBigEndian(d)
    FloatingPointEncoders-->>DoubleFloatHelper: Array[Byte]
    DoubleFloatHelper-->>EncoderSelector: Encoder
    EncoderSelector-->>ParserVisitor: Some(Encoder)

Loading

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~50 minutes

Possibly related PRs

• AbsaOSS/cobrix#783: Modifies the same EncoderSelector.scala encoder dispatch pipeline that this PR extends for COMP-1/COMP-2 support.

Poem

A bunny once hopped through mainframe land,
Where COMP-1 and COMP-2 values expand.
IEEE-754, IBM hex-float bytes in place,
Big-endian, little-endian—encoded with grace.
🐇 With a flip and a scale, the floats find their way,
Written to EBCDIC data, hooray! 🎉

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title directly and specifically describes the main change: adding COMP-1 and COMP-2 floating point encoders for the writer, which matches the core objective of the PR.
Linked Issues check	✅ Passed	The code changes successfully implement support for writing COMP-1 and COMP-2 fields by introducing floating point encoders with multiple format support (IEEE754, IBM, endianness variants) and updating the writer to use them.
Out of Scope Changes check	✅ Passed	All changes are focused on the floating point encoder implementation and writer support for COMP-1/COMP-2. No unrelated modifications detected outside the stated objectives.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/855-add-comp-1-2-support-for-writer

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[github-actions]

JaCoCo code coverage report - 'cobol-parser'

Overall Project	90.17% -0.04%	🍏
Files changed	96.27%	🍏

File	Coverage
FloatingPointEncoders.scala	97.35% -2.65%	🍏
ParserVisitor.scala	89.46%	🍏
EncoderSelector.scala	87.2% -2.4%	🍏

[github-actions]

JaCoCo code coverage report - 'spark-cobol'

Overall Project	83.4%	🍏

There is no coverage information present for the Files changed

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In
`@cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/EncoderSelector.scala`:
- Line 189: In the getDoubleFloatingPointEncoder method, the fallback case at
line 189 converts values using x.toString.toFloat which causes precision loss.
Change this to x.toString.toDouble to maintain proper double-precision accuracy
when handling non-standard input types in the encoder.

🪄 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: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 15ddd494-8ff3-45f9-a705-9e95a7a325c6

📥 Commits

Reviewing files that changed from the base of the PR and between 1e6e9f7 and 0e4a2d5.

📒 Files selected for processing (4)

• cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/antlr/ParserVisitor.scala
• cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/EncoderSelector.scala
• cobol-parser/src/main/scala/za/co/absa/cobrix/cobol/parser/encoding/FloatingPointEncoders.scala
• spark-cobol/src/test/scala/za/co/absa/cobrix/spark/cobol/writer/FixedLengthEbcdicWriterSuite.scala