feat: parse typed process outputs from the language server#12
Merged
Conversation
Make `parse_process_hover` extract outputs for processes written in the
typed Nextflow DSL (`nextflow.enable.types`). The language server emits
typed outputs as `channel: Record { field: Type ... }` (multi-line,
brace-delimited), which the previous parser did not recognize: each
output line returned `None`, `process.outputs` ended up empty, and the
renderer omitted the Outputs section entirely. Typed inputs already
worked and are unchanged.
Implementation:
* `_extract_output_block` (new) replaces the regex-only output-section
capture in `parse_process_hover`. It is brace-aware so a typed
`Record { ... }` is captured in full and multiple output channels are
not truncated at the first closing brace. The match is anchored to a
line start so an `output:` appearing mid-line (e.g. inside a script
string) is not mistaken for the section label, and any declaration
sharing the `output:` line is included as the first line of the block.
* `_parse_output_declarations` is now brace-aware and uses one-way
dataflow: when a typed-record line opens at depth 0 it is buffered in
`pending`; the record's direct field lines (depth 1) are collected;
the finalized `ParsedOutput` is constructed once on the closing `}`,
with the record's field shape surfaced in `name` (e.g.
`meta: Map, bam: Path`). Deeper-nested record lines are not expanded
and field lines never become separate outputs.
* `_parse_single_output` gains two typed branches: `name: Record { ...`
and the simple typed form `name: Type`. The simple branch is guarded
against `_TRADITIONAL_QUALIFIERS` so it cannot false-match traditional
declarations.
* `_typed_to_qualifier` maps `"Record"` to `"record"` (lowercase, for
consistency with `val`/`path`), keeping the type-name-to-label mapping
in one place. `_parse_output_declarations` keys off the parsed type
via the cached `_typed_to_qualifier("Record")` so the "is this a
record channel?" test has a single source of truth.
* The legacy `_OUTPUT_SECTION_RE` is removed: `enrich_outputs_from_source`
now uses the new brace-aware extractor too, so the hover and source
paths share one section extractor. This also fixes a latent
truncation when a source output glob contains brace expansion such as
`path "*_{R1,R2}.fastq"`.
Tests:
* Typed record outputs (single channel, multiple channels, mixed with
traditional outputs, nested `Record { Record { ... } }`).
* Typed simple outputs (`Path` -> `path`; `Integer` -> `val`).
* Field-collection edge cases: empty record falls back to emit-as-name;
a record-opening `{` followed by trailing whitespace does not leak
the brace into the field text.
* A declaration sharing the `output:` line is captured.
* `_extract_output_block` ignores an `output:` appearing mid-line and
returns `None` when no output section is present.
* `enrich_outputs_from_source` correctly handles a source whose output
glob contains brace expansion (regression for the old truncating
regex).
Owner
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Make
parse_process_hoverextract outputs for processes written in the typed Nextflow, 26.04 (nextflow.enable.types), with or without Records.Implementation:
_extract_output_block(new) replaces the regex-only output-section capture inparse_process_hover. It is brace-aware so a typedRecord { ... }is captured in full and multiple output channels are not truncated at the first closing brace. The match is anchored to a line start so anoutput:appearing mid-line (e.g. inside a script string) is not mistaken for the section label, and any declaration sharing theoutput:line is included as the first line of the block._parse_output_declarationsis now brace-aware with the record's field shape surfaced inname(e.g.meta: Map, bam: Path). Deeper-nested record lines are not expanded and field lines never become separate outputs._parse_single_outputgains two typed branches:name: Record { ...and the simple typed formname: Type. The simple branch is guarded against_TRADITIONAL_QUALIFIERSso it cannot false-match traditional declarations._typed_to_qualifiermaps"Record"to"record"(lowercase, for consistency withval/path), keeping the type-name-to-label mapping in one place._OUTPUT_SECTION_REis removed:enrich_outputs_from_sourcenow uses the new brace-aware extractor too, so the hover and source paths share one section extractor. This also fixes a latent truncation when a source output glob contains brace expansion such aspath "*_{R1,R2}.fastq".What it looks like:
examples.zip
Beyond a general review:
Claude and I did a lot of work coming up with "odd" patterns we might use. Did we catch them all? Did we cover cases that can't exist and we could simplify the code?
Current Tests:
Record { Record { ... } }).Path->path;Integer->val).{followed by trailing whitespace does not leak the brace into the field text.output:line is captured._extract_output_blockignores anoutput:appearing mid-line and returnsNonewhen no output section is present.enrich_outputs_from_sourcecorrectly handles a source whose output glob contains brace expansion (regression for the old truncating regex).