orxtra v0.13.0 /tool.src.orxtra.tool._data_tool_http
Updated
On this page

Builds a Tool from an HttpExecution definition: substitutes {{secret:NAME}} placeholders at call time, interpolates {param} in the URL, issues the httpx request, and validates the response against the output schema.

#tool.src.orxtra.tool._data_tool_http

#tool.src.orxtra.tool._data_tool_http

Factory for building Tool instances from HttpExecution definitions.

Takes a DataToolDefinition with an HttpExecution config and builds a concrete Tool whose execute function:

  1. Validates agent-supplied args against the param schema.
  2. Substitutes {{secret:NAME}} placeholders in URL, headers, and

body_template at CALL TIME (real values never enter the Tool object or LLM-visible text).

  1. Interpolates {param_name} in the URL with URL-encoded arg

values, validating against param patterns when defined.

  1. Makes the HTTP request via httpx.
  2. Validates the response body against the output schema (if defined).
  3. Returns a ToolOutput with validated/projected data and rendered text.

#_build_json_schema_params

python
def _build_json_schema_params(params: dict[str, ParamDef]) -> dict[str, Any]

Build a JSON Schema parameters dict from ParamDef entries.

Delegates to the shared implementation in _data_tool_shared.

#_substitute_secrets

python
def _substitute_secrets(text: str, secret_registry: SecretRegistry | None) -> str

Replace {{secret:NAME}} placeholders with real values.

Hard error if placeholders exist but no registry is provided.

#_interpolate_url

python
def _interpolate_url(url_template: str, args: dict[str, Any], params: dict[str, ParamDef]) -> str

Substitute {param_name} in the URL with URL-encoded arg values.

Validates interpolated param values against their pattern if one is defined. Raises ToolError on missing params or pattern mismatches.

#_validate_output_schema

python
def _validate_output_schema(response_data: Any, schema: dict[str, Any]) -> None

Validate response data against the output JSON Schema.

Delegates to the shared implementation in _data_tool_shared.

#_validate_args

python
def _validate_args(args: dict[str, Any], params: dict[str, ParamDef]) -> None

Validate that required params are present and types match.

Delegates to the shared implementation in _data_tool_shared.

#_apply_secret_substitution

python
def _apply_secret_substitution(url_template: str, header_template: dict[str, str] | None, body_tmpl: str | None, secret_registry: SecretRegistry | None) -> tuple[str, dict[str, str] | None, str | None]

Apply secret substitution to URL, headers, and body template.

Returns (effective_url, effective_headers, effective_body).

#_make_http_request

python
async def _make_http_request(*, method: str, url: str, headers: dict[str, str] | None, body: str | None, timeout_ceiling: int) -> tuple[httpx.Response, int]

Execute the HTTP request. Returns (response, elapsed_ms).

#build_http_tool

python
def build_http_tool(definition: DataToolDefinition, deps: ToolDeps, timeout_ceiling: int=30) -> Tool

Build a Tool from a DataToolDefinition with HttpExecution config.

Args:

  • definition: A validated DataToolDefinition with type = "http".
  • deps: Session-scoped dependencies (secret_registry,

preview_threshold, preview_lines).

  • timeout_ceiling: Maximum HTTP timeout in seconds.

Returns:

  • A Tool instance ready for execution pipeline wrapping.
Search