FlexiRule Docs

Resolver

Resolvers are the "Data Glue" of FlexiRule. They solve a fundamental problem: **How do we get dynamic data into an action without writing Python?**

Resolver: Resolver Engine#

Resolvers are the “Data Glue” of FlexiRule. They solve a fundamental problem: How do we get dynamic data into an action without writing Python?

Instead of hardcoding field paths, FlexiRule uses a Value Resolution Pipeline.

The engine uses a Compiled Resolver Graph. When a rule is loaded, its dynamic values are “compiled” into optimized resolver objects.

StaticResolver: Returns a constant value.
VariableResolver: Resolves a path like doc.customer_name or vars.total.
MathFormulaResolver: Performs arithmetic (a + b, a * b) on context variables.
DateFormulaResolver: Handles date math (Today + 7 days).
NormalizationResolver: Cleans data (Trimming, Scrubbing).
FormatResolver: Formats dates and currency.
ExpressionResolver: The most powerful—it chains multiple segments (Text + Token + Token) into a single result.

Request: An action needs a value (e.g., “Assign to Field”).
Dispatch: The ValueResolver.compile method determines the resolver type from the configuration JSON.
Execution: The resolve(context) method is called. It performs the logic (e.g., fetching from DB, calculating a sum).
Output: The raw value is returned to the action.

Because resolvers are objects, they can be nested.

A NormalizationResolver can take the output of a VariableResolver.
An ExpressionResolver can combine a StaticResolver (“Order #”) with a VariableResolver (doc.name).

This enables Low-Code Logic Composition:

“Format the (Sum of (Child Table.Amount)) as Currency (USD)”

The architecture is prepared for even more advanced patterns:

Recursive Resolution: Resolvers that resolve other resolvers.
Lazy Evaluation: Values that are only calculated when actually accessed.
Caching: The get_compiled_resolver helper already uses request-local caching, making the system extremely fast even with hundreds of dynamic bindings.