Actions14
Create Package
AI-generatedOverview
This node operation creates a new package (referred to as a 'prompt' in the API) within a specified playbook in the HumanFirst platform. It allows users to define various parameters for the package, such as its name, contents, natural language generation (NLG) model parameters (temperature, top-p, max tokens, stop sequences, frequency penalty, presence penalty), post-processing options, run mode, parent package ID, position, metadata, source information, additional parameters, and an optional NLG timeout. This operation is useful for automating the creation of customized packages in a playbook, which can be used for tasks like generating text completions or managing conversational AI workflows.
Use Case Examples
- Creating a new prompt package in a playbook with specific NLG parameters and metadata.
- Automating the addition of a package to a playbook with custom post-processing and run mode settings.
Properties
| Name | Meaning |
|---|---|
| Namespace | The namespace under which the playbook exists, required to identify the playbook context. |
| Playbook ID | The unique identifier of the playbook where the package will be created. |
| Name | The name of the package to be created. Defaults to 'Package' with a timestamp suffix if not changed. |
| Contents | The main content or body of the package. |
| Temperature | NLG model parameter controlling randomness in text generation. |
| Top-P | NLG model parameter controlling nucleus sampling for text generation. |
| Max Tokens | Maximum number of tokens to generate in the output. |
| Stop Sequences (JSON Array) | JSON array of strings that define stop sequences for the text generation. |
| Frequency Penalty | NLG model parameter to penalize new tokens based on their existing frequency in the text so far. |
| Presence Penalty | NLG model parameter to penalize new tokens based on whether they appear in the text so far. |
| Post-Processing | Option to specify how the generated text should be post-processed (e.g., none, newlines, conversations, key-value, delimiter). |
| Delimiter (if Post-Processing = Delimiter) | Custom delimiter string used if post-processing is set to 'Delimiter'. |
| Run Mode | Determines whether the package runs once merging all items or runs on each item individually. |
| Parent Package ID | Optional ID of a parent package to establish hierarchy. |
| Position | Position index of the package within the playbook. |
| Metadata (JSON) | Additional metadata for the package in JSON format. |
| Source (JSON) | Source information for the package in JSON format. |
| Parameters (JSON Array) | Array of parameters for the package in JSON format. |
| NLG Timeout (ISO Duration) | Optional timeout duration for natural language generation in ISO format. |
Output
JSON
json- The JSON response from the HumanFirst API containing details of the created package.
Dependencies
- HumanFirst API key credential required for authentication.
Troubleshooting
- Ensure that the 'namespace' and 'playbookId' parameters are correctly set and correspond to existing entities in the HumanFirst platform.
- Verify that JSON inputs such as 'stop_sequences', 'metadata', 'source', and 'parameters' are valid JSON strings to avoid parsing errors.
- If the package name is left as default 'Package', a timestamp suffix is automatically added to avoid naming conflicts.
- API request failures may occur due to network issues or invalid API keys; check credentials and network connectivity.
- Timeouts or errors during package creation might be related to large input sizes or server-side issues; consider adjusting the 'NLG Timeout' parameter or retrying later.