Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/tudoumono/Sherpa/llms.txt

Use this file to discover all available pages before exploring further.

Impact analysis traces which modules, copybooks, data items, batch jobs, and other artefacts are affected when a given knowledge-graph node changes. You supply a starting term (a graph node name, an alias, or a business-rule label), and Sherpa resolves it against the world’s Neo4j graph, traverses impact relationships up to eight hops, and returns a structured list of affected items with judgements, evidence, and traversal paths. Results can be exported to Excel for offline review.

POST /impact/run

POST /impact/run Resolve a starting node in the knowledge graph and compute all items that depend on it. The analysis runs synchronously and the result is stored in memory under the returned analysis_id.

Request body

start
string
required
The name of the graph node to use as the impact origin — a data item, module name, business rule label, or any alias recognised by the world’s alias map.
version
string
default:"v1"
World identifier (the registered document folder). Must match ^[A-Za-z0-9][A-Za-z0-9._-]{0,63}$.
include_deprecated
boolean
default:"false"
When false (default), deprecated and hidden-candidate nodes and relationships are excluded from the traversal. Set to true to include them.
scope_paths
string[]
default:"[]"
Restrict the traversal to subfolders whose paths start with one of these prefixes. An empty list searches the entire world. Unknown prefixes return 422.

Response

analysis_id
integer
Unique ID for this analysis result. Use it with GET /impact/{aid} and GET /impact/{aid}/export.xlsx.
status
string
Always "done" for a successful synchronous run.
count
integer
Total number of affected items found.
sure
integer
Number of items with a "sure" judgement (all graph edges are statically extracted).
review
integer
Number of items with a "review" judgement (at least one LLM-extracted edge in the path).
presumed
integer
Number of items inferred via full-text search when no direct graph path was found (judgement: "presumed").
curl -s -X POST http://localhost:8000/impact/run \
  -H "Content-Type: application/json" \
  -b "sherpa_session=<token>" \
  -d '{
    "start": "CALC-TAX",
    "version": "payroll-2024",
    "include_deprecated": false,
    "scope_paths": ["src/tax"]
  }'

GET /impact/

GET /impact/{aid} Retrieve the full result of a previously run impact analysis. Results are held in memory for the lifetime of the server process.

Path parameters

aid
integer
required
Analysis ID returned by POST /impact/run.

Response

type
string
Always "impact".
world_id
string
The world the analysis was run against.
start
string
The starting node name as supplied to the request.
include_deprecated
boolean
Whether deprecated nodes were included.
scope_prefixes
string[]
The normalised scope prefix list used during traversal.
starts
object[]
The resolved starting nodes in the graph.
items
object[]
Sorted list of impacted artefacts. Items with judgement: "sure" come first, then sorted by category and name.
presumed
object[]
Items inferred via full-text co-occurrence when no direct graph path was found. Same shape as items, always with judgement: "presumed".
curl -s http://localhost:8000/impact/1 \
  -b "sherpa_session=<token>"

GET /impact//export.xlsx

GET /impact/{aid}/export.xlsx Download the impact analysis as an Excel workbook. The file is generated in the current user’s personal workspace outputs/ directory and served as a file attachment. This action is recorded in the audit log.

Path parameters

aid
integer
required
Analysis ID returned by POST /impact/run.

Response

Binary Excel file (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet) with a Content-Disposition: attachment; filename="impact_<aid>.xlsx" header.
Analysis results are stored in memory only. If the server restarts, previously issued analysis_id values will no longer be resolvable and return 404.
curl -s -OJ http://localhost:8000/impact/1/export.xlsx \
  -b "sherpa_session=<token>"

Build docs developers (and LLMs) love