Support bibliography citeproc tests #273

PgBiel · 2025-02-03T20:04:41Z

Fixes #229.

As a drive-by, affixes around bibliography <layout> are now included in the output. (This should be the only visible change of this PR to users.)

Initial inspiration was taken from #228, and includes its drive-by fix of affixes in bibliography layout, but the approach taken for test comparison is entirely different: I implemented code to generate HTML similar to that generated by citeproc for the citeproc tests. The generated HTML is not indented, and so indentation in the original test result is stripped for comparison.

Future work

This revealed a bug, it appears that formatting is always being applied to affixes by hayagriva (cf sort_VariousNameMacros1.txt, where bold is being applied to spaces around a bold group), but this can be fixed later.

Also, there are various test failures to investigate, for example "bugreports_parenthesis seems to have some extraneous spacing to the left on hayagriva's output, which is weird. Perhaps citeproc is compressing two consecutive spaces in a way we didn't think of.

In addition, I considered moving affix application to layout::render to apply it to anywhere else using bibliography, but it appears this is already manually handled by citation, so this breaks things.

Scripts used

Nushell file (parse.nu):

# Converts citeproc test in stdin into record with one column per section:
# { MODE: "citation | bibliography | ...", RESULT: "...", ... }
def parse-citeproc-test [] {
  parse --regex '(?:>>=+ (?P<section>[\w\-]+) =+>>
(?P<body>[\s\S]+?)
<<=+ [\w\-]+ =+<<)' | transpose --header-row -d
}

# Parse all tests in a folder (or current folder)
# Returns a table with columns name (test filename), MODE (citation / bibliography), RESULT, ...
def parse-citeproc-test-dir [dir: string = "."] {
  ls $dir | filter { $in.type == "file" } | get name | each { |name| { name: $name, results: ($"(open --raw $name)" | parse-citeproc-test) } } | flatten
}

# Converts parsed XML or HTML (from '$row.RESULT | from xml', which is a good approximation for our case)
# into a flattened list of element tag / attribute / text information. 
# Good to extract all kinds of elements.
def html-flattener []: record<tag: string, attributes: record, content: any> -> list<record<tag: string, attributes: record, text: any>> {
  let rec = $in
  mut queue = [$rec]
  mut results = []
  while $queue != [] {
    let next = $queue | last
    $queue = $queue | drop
    let desc = $next.content | describe

    let text = if $desc == "string" { $next.content }
    let attributes = if $next.attributes == null { {} } else { $next.attributes }
    $results = $results | append {tag: $next.tag, attributes: $attributes, text: $text}

    if $desc != "nothing" and $desc != "string" {
      # Content is a list of elements
      $queue = $queue | append $next.content
    }
  }

  return $results
}

# Receives a result from 'parse-citeproc-test-dir' and returns parsed bibliography tests only.
# The parsed_res column contains parsed HTML.
# The flattened_res column contains the flattened parsed HTML (from html-flattener), which is a list
# of element tag / attributes / text.
def parse-citeproc-bib-tests []: table -> table { 
  filter { $in.MODE? == "bibliography" }
  | insert parsed_res { if $in.RESULT? != null { $in.RESULT | from xml } }
  | insert flattened_res { if $in.parsed_res? == null { [] } else { get parsed_res | html-flattener } }
}

You can use this at processor-tests/humans with

source parse.nu
let parsed_bibs = parse-citeproc-test-dir | parse-citeproc-bib-tests

Which provides a table with columns name (filename), MODE (citation / bibliography), RESULT (test result), ..., as well as parsed_res (RESULT parsed as HTML (actually XML, but here that works) into a table) and flattened_res (flattened list of HTML element information)

The command below provides the list of HTML elements used:

$parsed_bibs.flattened_res | flatten | filter { $in.tag != null } | uniq

Output

╭────┬──────┬──────────────────────────────────────┬──────╮
│  # │ tag  │              attributes              │ text │
├────┼──────┼──────────────────────────────────────┼──────┤
│  0 │ div  │ ╭───────┬──────────────╮             │      │
│    │      │ │ class │ csl-bib-body │             │      │
│    │      │ ╰───────┴──────────────╯             │      │
│  1 │ div  │ ╭───────┬───────────╮                │      │
│    │      │ │ class │ csl-entry │                │      │
│    │      │ ╰───────┴───────────╯                │      │
│  2 │ i    │ {record 0 fields}                    │      │
│  3 │ span │ ╭───────┬──────────────────────────╮ │      │
│    │      │ │ style │ font-variant:small-caps; │ │      │
│    │      │ ╰───────┴──────────────────────────╯ │      │
│  4 │ div  │ ╭───────┬──────────────────╮         │      │
│    │      │ │ class │ csl-right-inline │         │      │
│    │      │ ╰───────┴──────────────────╯         │      │
│  5 │ b    │ {record 0 fields}                    │      │
│  6 │ div  │ ╭───────┬─────────────────╮          │      │
│    │      │ │ class │ csl-left-margin │          │      │
│    │      │ ╰───────┴─────────────────╯          │      │
│  7 │ div  │ ╭───────┬───────────╮                │      │
│    │      │ │ class │ csl-block │                │      │
│    │      │ ╰───────┴───────────╯                │      │
│  8 │ sup  │ {record 0 fields}                    │      │
│  9 │ span │ ╭───────┬──────────╮                 │      │
│    │      │ │ style │ baseline │                 │      │
│    │      │ ╰───────┴──────────╯                 │      │
│ 10 │ div  │ ╭───────┬────────────╮               │      │
│    │      │ │ class │ csl-indent │               │      │
│    │      │ ╰───────┴────────────╯               │      │
│ 11 │ span │ ╭───────┬────────────────────╮       │      │
│    │      │ │ style │ font-style:normal; │       │      │
│    │      │ ╰───────┴────────────────────╯       │      │
│ 12 │ span │ ╭───────┬──────────────────────╮     │      │
│    │      │ │ style │ font-variant:normal; │     │      │
│    │      │ ╰───────┴──────────────────────╯     │      │
│ 13 │ sub  │ {record 0 fields}                    │      │
╰────┴──────┴──────────────────────────────────────┴──────╯

TODO

Fix some bugs and differences on some failing tests (e.g. no formatting when the text is just whitespace...)
- Actually, this formatting difference (from the sort_VariousNameMacros1.txt test) is probably a bug in hayagriva and best fixed separately. I think the problem is that formatting is being applied to affixes when it shouldn't.
Share some of the scripts and commands I used perhaps

(TODO: bugreports_parenthesis should work... is citeproc trimming things?)

tests/citeproc.rs

PgBiel and others added 7 commits February 3, 2025 16:48

initial bib test support

7e27534

Fix bib layout affixes

76c0c5c

remove tests which aren't passing yet

652099c

(TODO: bugreports_parenthesis should work... is citeproc trimming things?)

add missing passing test with affixes fix

7e46bfc

move regex to a dev dependency

7673edf

format

5b82742

clippy

355ebd7

reknih reviewed Feb 4, 2025

View reviewed changes

tests/citeproc.rs Outdated Show resolved Hide resolved

tests/citeproc.rs Outdated Show resolved Hide resolved

tests/citeproc.rs Outdated Show resolved Hide resolved

PgBiel added 5 commits February 4, 2025 16:11

fix second-field-align

4255751

do not output anchor elements for links

771da69

adjust comments

3550d04

adjust missing test

415e400

use 'apply_(affix)' for bibliography layout affixes

f5880dd

PgBiel marked this pull request as ready for review February 4, 2025 23:59

PgBiel merged commit 1058be7 into main Feb 5, 2025
4 checks passed

PgBiel deleted the bib-citeproc-tests branch February 5, 2025 01:09

reknih mentioned this pull request Feb 5, 2025

Add support for subsequent author substitution #228

Open

5 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Support bibliography citeproc tests #273

Support bibliography citeproc tests #273

PgBiel commented Feb 3, 2025 •

edited

Loading

Support bibliography citeproc tests #273

Support bibliography citeproc tests #273

Conversation

PgBiel commented Feb 3, 2025 • edited Loading

Future work

Scripts used

TODO

PgBiel commented Feb 3, 2025 •

edited

Loading