refactor(api): Separate SegmentType for Integer/Float to Enable Pydantic Serialization (#22025)
refactor(api): Separate SegmentType for Integer/Float to Enable Pydantic Serialization (#22025)
This PR addresses serialization issues in the VariablePool model by separating the `value_type` tags for `IntegerSegment`/`FloatSegment` and `IntegerVariable`/`FloatVariable`. Previously, both Integer and Float types shared the same `SegmentType.NUMBER` tag, causing conflicts during serialization.
Key changes:
- Introduce distinct `value_type` tags for Integer and Float segments/variables
- Add `VariableUnion` and `SegmentUnion` types for proper type discrimination
- Leverage Pydantic's discriminated union feature for seamless serialization/deserialization
- Enable accurate serialization of data structures containing these types
Closes #22024.
feat: Persist Variables for Enhanced Debugging Workflow (#20699)
This pull request introduces a feature aimed at improving the debugging experience during workflow editing. With the addition of variable persistence, the system will automatically retain the output variables from previously executed nodes. These persisted variables can then be reused when debugging subsequent nodes, eliminating the need for repetitive manual input.
By streamlining this aspect of the workflow, the feature minimizes user errors and significantly reduces debugging effort, offering a smoother and more efficient experience.
Key highlights of this change:
- Automatic persistence of output variables for executed nodes.
- Reuse of persisted variables to simplify input steps for nodes requiring them (e.g., `code`, `template`, `variable_assigner`).
- Enhanced debugging experience with reduced friction.
Closes #19735.
feat(api): Adjust `WorkflowDraftVariable` and `WorkflowNodeExecutionModel` (#20746)
- Add `node_execution_id` column to `WorkflowDraftVariable`, allowing efficient implementation of
the "Reset to last run value" feature.
- Add additional index for `WorkflowNodeExecutionModel` to improve the performance of last run lookup.
Closes #20745.
Simplify `execution_metadata` Handling for `WorkflowNodeExecution` (#20062)
Currently, `WorkflowNodeExecution.execution_metadata_dict` returns `None` when metadata is absent in the database. This requires all callers to perform `None` checks when processing metadata, leading to more complex caller-side logic.
This pull request updates the `execution_metadata_dict` method to return an empty dictionary instead of `None` when metadata is absent. This change would simplify the caller logic, as it removes the need for explicit `None` checks and provides a more consistent data structure to work with.
feat(api): Introduce `WorkflowDraftVariable` Model (#19737)
- Introduce `WorkflowDraftVariable` model and the corresponding migration.
- Implement `EnumText`, a custom column type for SQLAlchemy designed
to work seamlessly with enumeration classes based on `StrEnum`.
Alembic's offline mode generates SQL from SQLAlchemy migration operations,
providing developers with a clear view of database schema changes without
requiring an active database connection.
However, some migration versions (specifically bbadea11be and d7999dfa4a)
were performing database schema introspection, which fails in offline mode
since it requires an actual database connection.
This commit:
- Adds offline mode support by detecting context.is_offline_mode()
- Skips introspection steps when in offline mode
- Adds warning messages in SQL output to inform users that assumptions were made
- Prompts users to review the generated SQL for accuracy
These changes ensure migrations work consistently in both online and offline modes.
Close #19284.
feat(api): Add image multimodal support for LLMNode (#17372)
Enhance `LLMNode` with multimodal capability, introducing support for
image outputs.
This implementation extracts base64-encoded images from LLM responses,
saves them to the storage service, and records the file metadata in the
`ToolFile` table. In conversations, these images are rendered as
markdown-based inline images.
Additionally, the images are included in the LLMNode's output as
file variables, enabling subsequent nodes in the workflow to utilize them.
To integrate file outputs into workflows, adjustments to the frontend code
are necessary.
For multimodal output functionality, updates to related model configurations
are required. Currently, this capability has been applied exclusively to
Google's Gemini models.
Close #15814.
Signed-off-by: -LAN- <laipz8200@outlook.com>
Co-authored-by: -LAN- <laipz8200@outlook.com>
Junyan Qin (Chin)
Maries
Asuka Minato
Kushagra Singhal
znn
Aryan Raj
-LAN-
yihong
QuantumGhost
Novice
NeatGuyCoding
jiangbo721
YoungLH
Qiang Lee
Xin Zhang
Good Wood
非法操作
湛露先生
Ganondorf
crazywoola
sondin
Will
Jyong