Use pattern string instead of compiled pattern in field_specs.py #967
Conversation
… the compiled pattern. This required removing the lookahead portion of the string, but because we're using min_length and max_length in the spec it shouldn't matter.
ahouseholder
added
tech/backend
sei-vsarvepalli
left a comment
sei-vsarvepalli
left a comment
There was a problem hiding this comment.
Well.. the tests are running okay. This regex is one I cannot read :)
|
The only change to the regex is that we're no longer checking min/max length within the pattern itself. This is okay because both pydantic and json schema have length enforcement via first-order properties that we're already using. So the pattern doesn't need to care about length anyway. This was discovered because when I changed Also, this was the only place where we were using a compiled pattern in creating a pydantic object. Other places where we used patterns we were always including the pattern string instead, so this makes us consistent throughout. |
|
What about using NS_PATTERN_STR instead of NS_PATTERN.pattern and remove NS_PATTERN entirely? It would be an additional simplification.
If you look at the different parts of the regex in pattern.py, that is close to the ABNF and can be understood. The one in NS_PATTERN_STR is just unrolled. ;) |
Hi Bernhard - I think ABNF is helpful. We are not pursuing much more update on it at this time. |
field_spec.py#965This PR replaces the compiled NS_PATTERN with the pattern string when creating the field specs.
A side effect of this change is that the pattern no longer needs the lookahead for length check, as we're specifying as part of the pydantic class and that carries into the JSON schema directly without it having to be part of the regex.
Copilot Summary
This pull request simplifies and standardizes the namespace pattern validation logic across several schema and utility files. The main change is the removal of an unnecessary length check from the regular expression pattern, relying instead on explicit
minLengthandmaxLengthconstraints. This makes the regex easier to maintain and ensures consistency between schema validation and Python-side checks.Most important changes:
Schema pattern updates:
(?=.{3,1000}$)) from thepatternfield in the JSON schema definitions forDecisionPoint_2_0_0.schema.json,DecisionPointGroup_2_0_0.schema.json,DecisionTable_2_0_0.schema.json,SelectionList_2_0_0.schema.json, andSsvcObjectRegistry_2_0_0.schema.json, so that the regex only enforces the namespace format, not its length. Length is now enforced solely by theminLengthandmaxLengthschema properties. [1] [2] [3] [4] [5] [6] [7]Python utility and test consistency:
src/ssvc/utils/patterns.pyto remove the length check from the pattern string, matching the schema change.src/ssvc/utils/field_specs.pyto use the raw pattern string from the compiled regex, ensuring consistency between the schema and Python validation.src/test/test_namespaces.pyto accumulate and report all unexpected matches, making debugging easier.