Guides
Python SDK
from dev_guru import GuruClient
with GuruClient(workspace="path/to/project") as guru:
print(guru.doctor().compatible) # True when Ananke is in range
result = guru.explore(name="My project") # Ananke builds the graph; Dev Guru projects it
page = guru.export(result, output="reports/explorer.html")
print(result.model["project"]) # {'files': 25, 'symbols': 146, 'imports': 26, 'calls': 88, ...}
print(result.manifest["input_digest"])| Call | Returns |
|---|---|
GuruClient(workspace=".", transport="in-process") | a client bound to one explicit transport |
.doctor() | HarnessInfo: Ananke version, supported range, capabilities, reasons if incompatible |
.explore(name=None) | ExploreResult: the view model (dev-guru/explorer@1), its manifest and the transport |
.export(result, output=..., format="html") | the path of the written report (a manifest goes next to it) |
Errors
Failures raise dev_guru.errors.GuruError with a stable code, so branch on the code rather than on the message:
| Code | When |
|---|---|
HARNESS_MISSING | Ananke Plexus isn’t installed |
HARNESS_INCOMPATIBLE | its version is outside the supported range |
CAPABILITY_MISSING | it lacks a capability Dev Guru needs |
UNSUPPORTED_TRANSPORT | you asked for a transport that doesn’t exist (no silent fallback) |
WORKSPACE_INVALID | the folder doesn’t exist, or holds no Python files |
HARNESS_FAILED | Ananke reported success but stored nothing |
The lower-level port
dev_guru.client.connect() returns the AnankeClient port directly: info(), graph_data(workspace) (the graph
as plain data) and graph_summary(workspace). Everything Dev Guru knows about your code comes through this one port.