Skip to content

Some branch cleanup - #348

Merged
dwolfson merged 5 commits into
odpi:mainfrom
dwolfson:main
Sep 8, 2026
Merged

Some branch cleanup#348
dwolfson merged 5 commits into
odpi:mainfrom
dwolfson:main

Conversation

@dwolfson

@dwolfson dwolfson commented Sep 8, 2026

Copy link
Copy Markdown
Member

No description provided.

dwolfson and others added 5 commits September 6, 2026 17:12
pyproject.toml declares mcp >=0.1 but pyegeria/core/mcp_server.py imports
mcp.server.mcpserver.MCPServer, which requires mcp>=2.0 (confirmed live:
this dev venv has mcp==2.0.0 installed and the import works; the module
path doesn't exist in the 0.x/1.x mcp package line). Found by Egeria
Advisor rebuilding its demo deployment against pyegeria 6.1.10 -- worked
fine because they pinned mcp==2.1.1 explicitly, not because pyegeria's
own floor would have caught an older mcp.

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
…spec

format_set_executor.py's `token` parameter (ISSUE-86 — let a caller that
already holds an Egeria bearer token run a report without a password) is on
main with no dedicated test.

These tests existed only on an unpushed local branch,
claude/cranky-chebyshev-be2494, alongside a second, parallel implementation
of the same feature that main had meanwhile gained by another route.
Pushing that branch would have added a duplicate implementation; the tests
were the part worth keeping, so they are salvaged here on their own.

They were written against that branch's `_authenticate_client()` helper,
which main does not have — verified they pass unchanged against main's
inline implementation (7 passed), so they test the behaviour rather than
that refactor's shape.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
commands_project_compact.json's `Project Type` attribute declares six
valid_values (Project, Campaign, Task, PersonalProject, StudyProject,
Experiment) and describes four — `Project`, which is also the attribute's own
default_value, and `Experiment` are both missing from the prose.

The machine-readable list is correct: it matches Egeria 6's Project
classifications in OpenMetadataType.java (model 0130). Only the description is
behind.

It has propagated. gen_md_cmd_templates and gen_dr_help render the description,
so 25 files in the main tree carry the sentence verbatim — the source JSON plus
24 generated artifacts, 16 of which ship as user-facing templates. Notably
sample-data/templates/{basic,advanced}/Projects/Create_Experiment.md tells the
reader Experiment is not a supported value, and the generated help tables
contradict themselves in a single row by printing the prose beside the
valid_values column.

Logged, not fixed, per the standing rule for this repo. The entry carries a
suggested replacement description worded from the Egeria type definitions
themselves, the regenerate-and-propagate steps, and a note to check whether any
other compact-command attribute enumerates its valid_values in prose — the same
latent defect wherever it exists.

Found while mapping Resource Explorer's investigation classifications onto
Egeria's.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
@dwolfson
dwolfson merged commit 33c8a71 into odpi:main Sep 8, 2026
2 checks passed
pcoldico pushed a commit to pcoldico/egeria-python that referenced this pull request Sep 8, 2026
* docs(PYEGERIA_ISSUES): log ISSUE-91 -- mcp dependency floor too loose

pyproject.toml declares mcp >=0.1 but pyegeria/core/mcp_server.py imports
mcp.server.mcpserver.MCPServer, which requires mcp>=2.0 (confirmed live:
this dev venv has mcp==2.0.0 installed and the import works; the module
path doesn't exist in the 0.x/1.x mcp package line). Found by Egeria
Advisor rebuilding its demo deployment against pyegeria 6.1.10 -- worked
fine because they pinned mcp==2.1.1 explicitly, not because pyegeria's
own floor would have caught an older mcp.

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>

* test(view): cover ISSUE-86's caller-held bearer token in exec_report_spec

format_set_executor.py's `token` parameter (ISSUE-86 — let a caller that
already holds an Egeria bearer token run a report without a password) is on
main with no dedicated test.

These tests existed only on an unpushed local branch,
claude/cranky-chebyshev-be2494, alongside a second, parallel implementation
of the same feature that main had meanwhile gained by another route.
Pushing that branch would have added a duplicate implementation; the tests
were the part worth keeping, so they are salvaged here on their own.

They were written against that branch's `_authenticate_client()` helper,
which main does not have — verified they pass unchanged against main's
inline implementation (7 passed), so they test the behaviour rather than
that refactor's shape.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>

* ISSUE-92: Project Type's description lists 4 of its 6 valid_values

commands_project_compact.json's `Project Type` attribute declares six
valid_values (Project, Campaign, Task, PersonalProject, StudyProject,
Experiment) and describes four — `Project`, which is also the attribute's own
default_value, and `Experiment` are both missing from the prose.

The machine-readable list is correct: it matches Egeria 6's Project
classifications in OpenMetadataType.java (model 0130). Only the description is
behind.

It has propagated. gen_md_cmd_templates and gen_dr_help render the description,
so 25 files in the main tree carry the sentence verbatim — the source JSON plus
24 generated artifacts, 16 of which ship as user-facing templates. Notably
sample-data/templates/{basic,advanced}/Projects/Create_Experiment.md tells the
reader Experiment is not a supported value, and the generated help tables
contradict themselves in a single row by printing the prose beside the
valid_values column.

Logged, not fixed, per the standing rule for this repo. The entry carries a
suggested replacement description worded from the Egeria type definitions
themselves, the regenerate-and-propagate steps, and a note to check whether any
other compact-command attribute enumerates its valid_values in prose — the same
latent defect wherever it exists.

Found while mapping Resource Explorer's investigation classifications onto
Egeria's.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>

---------

Signed-off-by: Dan Wolfson <dan.wolfson@pdr-associates.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant