Skip to main content
pipeline_stats reports how much structure the ParaLens engine reconstructed from the transaction trace. This is primarily useful for advanced users, researchers, and debugging — it is safe to ignore in most consumer UIs.

​
Example

​
Fields

​
number
required
Count of normalized value movements reconstructed from the trace. Each entry represents a discrete transfer of economic value between addresses.
​
number
required
Count of non-value actions observed — approvals, event emissions, and other non-economic state changes that were processed but did not contribute directly to value flow.
​
number
required
Number of execution trace call frames inspected. Reflects the depth and breadth of the on-chain call graph walked during analysis.
​
number
required
Number of log entries (EVM events) inspected during reconstruction.
​
number
required
Number of nodes in the Token-Value Flow Graph (TVFG) — each node represents a unique address involved in value flow.
​
number
required
Number of directed edges in the Token-Value Flow Graph — each edge represents a value movement between two addresses.
​
number
required
Number of structural scopes discovered. Scopes are coherent sub-units of the trace, analogous to logical sub-transactions within the overall call tree.
​
number
required
Number of structural primitives discovered. Primitives are the lowest-level recognized patterns from which motifs and higher-order classifications are assembled.

​
Interpreting the values

Use the following heuristics when inspecting pipeline_stats for research or debugging purposes:

​
Usage guidance

Hide pipeline_stats in consumer-facing UIs. Surface it in analytics tools, analyst dashboards, or debug views where understanding engine reconstruction depth adds value.
  • Consumer UIs — omit this field entirely. It provides no actionable information for end users interpreting a transaction.
  • Analytics and research — use frames, tvfg_nodes, and tvfg_edges as complexity proxies when bucketing or filtering transactions by structural complexity.
  • Debugging — if a TxReport produces unexpected classifications, inspect pipeline_stats to confirm the engine successfully reconstructed the expected number of scopes and primitives.