feature(View): enabling view in SemanticLayerSchema

[scaliseraoul]

• Tests added and passed if fixing a bug or adding a new feature.
• All code checks passed.

---

Important

Adds view support to SemanticLayerSchema with new constraints, validation logic, and documentation updates.

• Behavior:
  • Adds view support in SemanticLayerSchema with constraints for column and relation formats.
  • Enforces mutual exclusivity between table and view in Source.
  • Validates column uniqueness and format in views.
• Models:
  • Adds Relation class to define relationships in views.
  • Updates Source class to include view attribute.
• Validation:
  • Adds check_columns_relations() in SemanticLayerSchema for view-specific validation.
  • Updates validate_type_and_fields() in Source for view constraints.
• Documentation:
  • Updates semantic-layer.mdx with view configuration and constraints.
• Tests:
  • Adds tests for view-specific validation in test_semantic_layer_schema.py.

^{This description was created by for cf142de. It will automatically update as commits are pushed.}

[ellipsis-dev]

❌ Changes requested. Reviewed everything up to cf142de in 2 minutes and 14 seconds

More details

• Looked at 338 lines of code in 3 files
• Skipped 0 files when reviewing.
• Skipped posting 8 drafted comments based on config settings.

1. tests/unit_tests/dataframe/test_semantic_layer_schema.py:113

• Draft comment:
  The view field should be a boolean value, not a string. Use True instead of 'true'.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable:
  While technically the comment is correct that True would be more proper than "true", the code is working as intended since the validation layer handles the type conversion. The test passes with the current implementation. Making this change would be a minor style improvement but doesn't affect functionality.
  The comment points out a real style issue. Using proper boolean types is generally better practice than string representations of booleans.
  However, since the validation layer already handles this gracefully and the tests pass, this is a very minor style issue that doesn't impact functionality or readability significantly.
  Delete the comment. While technically correct, it's a minor style suggestion that doesn't materially improve the code, and the current implementation works correctly.

2. pandasai/data_loader/semantic_layer_schema.py:93

• Draft comment:
  The error message here can be improved for clarity. Consider using:

                raise ValueError("A schema can only define either 'table' or 'view', not both.")

• Reason this comment was not posted:
  Marked as duplicate.

3. pandasai/data_loader/semantic_layer_schema.py:95

• Draft comment:
  The error message here can be improved for clarity. Consider using:

                raise ValueError("A schema must define either 'table' or 'view'.")

• Reason this comment was not posted:
  Marked as duplicate.

4. pandasai/data_loader/semantic_layer_schema.py:156

• Draft comment:
  The error message here can be improved for clarity. Consider using:

            raise ValueError("Column names must be unique. Duplicate column names detected.")

• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable:
  The original error message is already clear and grammatically correct. The suggested change is extremely minor - changing "names found" to "column names detected". This doesn't make the message meaningfully clearer or fix any actual problems. The suggestion feels like personal preference rather than a meaningful improvement.
  Perhaps in some contexts, being explicit about "column names" rather than just "names" could prevent ambiguity. The word "detected" may be more precise than "found".
  While those points are valid, the original message is already sufficiently clear in context - we're clearly talking about column names from the surrounding code, and "found" vs "detected" is purely stylistic.
  This comment should be removed as it suggests an extremely minor rewording that doesn't meaningfully improve clarity or fix any actual issues with the error message.

5. pandasai/data_loader/semantic_layer_schema.py:175

• Draft comment:
  The error message here can be improved for clarity. Consider using:

                    "All column names in a view must follow the format '[table].[column]'."

• Reason this comment was not posted:
  Marked as duplicate.

6. pandasai/data_loader/semantic_layer_schema.py:183

• Draft comment:
  The error message here can be improved for clarity. Consider using:

                    "All 'from' and 'to' parameters in relations must follow the format '[table].[column]'."

• Reason this comment was not posted:
  Marked as duplicate.

7. pandasai/data_loader/semantic_layer_schema.py:191

• Draft comment:
  The error message here can be improved for clarity. Consider using:

                    f"No relations defined for the following tables: {uncovered_tables}."

• Reason this comment was not posted:
  Marked as duplicate.

8. pandasai/data_loader/semantic_layer_schema.py:195

• Draft comment:
  The error message here can be improved for clarity. Consider using:

            "All column names in a table must not include a table prefix."

• Reason this comment was not posted:
  Marked as duplicate.

Workflow ID: wflow_GNrarDy3W8RIJwod

---

Want Ellipsis to fix these issues? Tag @ellipsis-dev in a comment. You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[ellipsis-dev]

The regex ^[a-zA-Z_]+\.[a-zA-Z_]+$ does not account for numeric characters in table or column names. Consider updating it to ^[a-zA-Z0-9_]+\.[a-zA-Z0-9_]+$ to allow numeric characters.

[scaliseraoul]

@ellipsis-dev can you commit the changes?

[gventuri]

Very well documented 💯