oxedyne/fe2o3/fe2o3_austenite/src/bib.rs
50.3 KiB, 58 runs
created by r1870400018:36623, which is this file's identity for as long as the history lasts, whatever it is later renamed to
download · who wrote it · its history
| 1 | //! Bibliography: a BibTeX reader and a Chicago author-date citation formatter. |
| 2 | //! |
| 3 | //! The engine reads a `.bib` file into [`Bibliography`], formats an in-text citation as |
| 4 | //! `(Surname Year)` through [`Bibliography::cite`], and sets a sorted reference list through |
| 5 | //! [`Bibliography::reference_list`]. It is standalone: the reader wires `#cite(<key>)` to `cite` |
| 6 | //! and the back-matter Bibliography section to `reference_list`; nothing here touches the parser or |
| 7 | //! the page. |
| 8 | //! |
| 9 | //! The oracle is Typst 0.15.1's `chicago-author-date` style, matched against a render of the |
| 10 | //! Lucronics bibliography. Three behaviours are Typst-specific and reproduced deliberately, each |
| 11 | //! noted at its site: the sort orders works by one author before works by that author with |
| 12 | //! coauthors (author count is the tie-break after the first surname); a journal article with no |
| 13 | //! volume sets a bare comma after the journal name (`*Journal*,.`); and page ranges are abbreviated |
| 14 | //! by the Chicago rule (`263--291` sets as `263-91`). |
| 15 | //! |
| 16 | //! [Written with AI entirely](https://need2know.ai/entirely-ai/code)\ |
| 17 | //! Anthropic Claude |
| 18 | |
| 19 | use oxedyne_fe2o3_core::prelude::*; |
| 20 | |
| 21 | use std::collections::BTreeMap; |
| 22 | |
| 23 | /// The closed set of entry types the formatter styles. An unrecognised `@type` reads as |
| 24 | /// [`EntryKind::Misc`], which sets title-italic like a book -- a declared fallback, not a hidden gap. |
| 25 | #[derive(Clone, Copy, Debug, PartialEq, Eq)] |
| 26 | pub enum EntryKind { |
| 27 | Book, |
| 28 | Article, |
| 29 | InCollection, |
| 30 | InProceedings, |
| 31 | TechReport, |
| 32 | Report, |
| 33 | Online, |
| 34 | Unpublished, |
| 35 | Misc, |
| 36 | } |
| 37 | |
| 38 | impl EntryKind { |
| 39 | fn from_type(s: &str) -> Self { |
| 40 | match s.to_ascii_lowercase().as_str() { |
| 41 | "book" | "booklet" | "proceedings" => Self::Book, |
| 42 | "article" => Self::Article, |
| 43 | "incollection" | "inbook" => Self::InCollection, |
| 44 | "inproceedings" | "conference" => Self::InProceedings, |
| 45 | "techreport" => Self::TechReport, |
| 46 | "report" => Self::Report, |
| 47 | "online" | "electronic" | "misc" => Self::Online, |
| 48 | "unpublished" => Self::Unpublished, |
| 49 | _ => Self::Misc, |
| 50 | } |
| 51 | } |
| 52 | } |
| 53 | |
| 54 | /// One personal or corporate name. A corporate name carries its whole literal in `family` with an |
| 55 | /// empty `given`; a personal name is inverted to `family, given` for the list and cited by `family`. |
| 56 | #[derive(Clone, Debug, PartialEq, Eq)] |
| 57 | pub struct Name { |
| 58 | family: String, |
| 59 | given: String, |
| 60 | corporate: bool, |
| 61 | } |
| 62 | |
| 63 | impl Name { |
| 64 | pub fn family(&self) -> &str { &self.family } |
| 65 | pub fn given(&self) -> &str { &self.given } |
| 66 | pub fn is_corporate(&self) -> bool { self.corporate } |
| 67 | |
| 68 | /// Sets the name inverted for a reference list: `Family, Given` (personal) or the literal |
| 69 | /// (corporate). |
| 70 | fn inverted(&self) -> String { |
| 71 | if self.corporate || self.given.is_empty() { |
| 72 | self.family.clone() |
| 73 | } else { |
| 74 | fmt!("{}, {}", self.family, self.given) |
| 75 | } |
| 76 | } |
| 77 | |
| 78 | /// Sets the name in natural order for a coauthor position: `Given Family`. |
| 79 | fn natural(&self) -> String { |
| 80 | if self.corporate || self.given.is_empty() { |
| 81 | self.family.clone() |
| 82 | } else { |
| 83 | fmt!("{} {}", self.given, self.family) |
| 84 | } |
| 85 | } |
| 86 | } |
| 87 | |
| 88 | /// A parsed BibTeX entry. Fields are held decoded (LaTeX accents and escapes resolved) in file |
| 89 | /// order; a reader wanting a raw field asks by name through [`Entry::field`]. |
| 90 | #[derive(Clone, Debug)] |
| 91 | pub struct Entry { |
| 92 | key: String, |
| 93 | kind: EntryKind, |
| 94 | fields: Vec<(String, String)>, |
| 95 | } |
| 96 | |
| 97 | impl Entry { |
| 98 | pub fn key(&self) -> &str { &self.key } |
| 99 | pub fn kind(&self) -> EntryKind { self.kind } |
| 100 | |
| 101 | /// Returns the decoded value of a field by name, case-insensitively. |
| 102 | pub fn field(&self, name: &str) -> Option<&str> { |
| 103 | self.fields.iter() |
| 104 | .find(|(k, _)| k.eq_ignore_ascii_case(name)) |
| 105 | .map(|(_, v)| v.as_str()) |
| 106 | } |
| 107 | |
| 108 | /// The names credited for citation: authors, or editors when there is no author. |
| 109 | pub fn credited(&self) -> Vec<Name> { |
| 110 | match self.field("author") { |
| 111 | Some(a) => parse_names(a), |
| 112 | None => match self.field("editor") { |
| 113 | Some(e) => parse_names(e), |
| 114 | None => Vec::new(), |
| 115 | }, |
| 116 | } |
| 117 | } |
| 118 | |
| 119 | fn year(&self) -> Option<i64> { |
| 120 | self.field("year").and_then(|y| { |
| 121 | let digits: String = y.chars().filter(|c| c.is_ascii_digit()).collect(); |
| 122 | digits.parse::<i64>().ok() |
| 123 | }) |
| 124 | } |
| 125 | } |
| 126 | |
| 127 | /// A styled fragment of a reference line. The block layer sets an [`RefStyle::Italic`] run in |
| 128 | /// italic; everything else -- quotation marks, full stops, DOI URLs -- is literal text. |
| 129 | #[derive(Clone, Debug, PartialEq, Eq)] |
| 130 | pub struct RefRun { |
| 131 | pub text: String, |
| 132 | pub style: RefStyle, |
| 133 | } |
| 134 | |
| 135 | #[derive(Clone, Copy, Debug, PartialEq, Eq)] |
| 136 | pub enum RefStyle { |
| 137 | Normal, |
| 138 | Italic, |
| 139 | } |
| 140 | |
| 141 | /// One formatted reference: the entry key and its runs, ready for the block layer. |
| 142 | #[derive(Clone, Debug, PartialEq, Eq)] |
| 143 | pub struct Reference { |
| 144 | pub key: String, |
| 145 | pub runs: Vec<RefRun>, |
| 146 | } |
| 147 | |
| 148 | impl Reference { |
| 149 | /// The reference as one plain string, italics flattened away. For tests and diagnostics. |
| 150 | pub fn plain(&self) -> String { |
| 151 | let mut s = String::new(); |
| 152 | for r in &self.runs { |
| 153 | s.push_str(&r.text); |
| 154 | } |
| 155 | s |
| 156 | } |
| 157 | } |
| 158 | |
| 159 | /// A parsed bibliography and the record of which keys a document has cited. |
| 160 | #[derive(Debug)] |
| 161 | pub struct Bibliography { |
| 162 | entries: Vec<Entry>, // in file order |
| 163 | index: BTreeMap<String, usize>, // key -> position in `entries` |
| 164 | cited: Vec<String>, // cited keys, first-cite order, deduplicated |
| 165 | } |
| 166 | |
| 167 | impl Bibliography { |
| 168 | /// Reads a BibTeX source string into a bibliography. Comments, `@string`/`@preamble` blocks and |
| 169 | /// stray text between entries are skipped; an entry with a duplicate key keeps the first seen. |
| 170 | pub fn parse(src: &str) -> Outcome<Self> { |
| 171 | let chars: Vec<char> = src.chars().collect(); |
| 172 | let mut entries: Vec<Entry> = Vec::new(); |
| 173 | let mut index: BTreeMap<String, usize> = BTreeMap::new(); |
| 174 | let mut i = 0usize; |
| 175 | let n = chars.len(); |
| 176 | while i < n { |
| 177 | // Advance to the next entry opener. |
| 178 | if chars[i] != '@' { |
| 179 | i += 1; |
| 180 | continue; |
| 181 | } |
| 182 | let (entry_opt, next) = res!(parse_entry(&chars, i)); |
| 183 | i = next; |
| 184 | if let Some(entry) = entry_opt { |
| 185 | if !index.contains_key(&entry.key) { |
| 186 | index.insert(entry.key.clone(), entries.len()); |
| 187 | entries.push(entry); |
| 188 | } |
| 189 | } |
| 190 | } |
| 191 | Ok(Self { entries, index, cited: Vec::new() }) |
| 192 | } |
| 193 | |
| 194 | pub fn len(&self) -> usize { self.entries.len() } |
| 195 | pub fn is_empty(&self) -> bool { self.entries.is_empty() } |
| 196 | |
| 197 | /// Borrows an entry by key. |
| 198 | pub fn entry(&self, key: &str) -> Option<&Entry> { |
| 199 | self.index.get(key).map(|&p| &self.entries[p]) |
| 200 | } |
| 201 | |
| 202 | /// Records a key as cited, in first-cite order, for the reference list. A key unknown to the |
| 203 | /// bibliography is ignored here and reported by [`Bibliography::cite`] at formatting time. |
| 204 | pub fn mark_cited(&mut self, key: &str) { |
| 205 | if self.index.contains_key(key) && !self.cited.iter().any(|k| k == key) { |
| 206 | self.cited.push(key.to_string()); |
| 207 | } |
| 208 | } |
| 209 | |
| 210 | /// Formats one in-text citation for a run of keys and records each as cited. Sets |
| 211 | /// `(Surname Year)` for one key and `(A Year; B Year)` for several, matching the oracle's |
| 212 | /// semicolon separator. |
| 213 | pub fn cite(&mut self, keys: &[&str]) -> Outcome<String> { |
| 214 | for &k in keys { |
| 215 | self.mark_cited(k); |
| 216 | } |
| 217 | self.format_citation(keys) |
| 218 | } |
| 219 | |
| 220 | /// Formats an in-text citation without recording it. Used where the reader has already marked |
| 221 | /// the keys, or for a pure formatting query. |
| 222 | pub fn format_citation(&self, keys: &[&str]) -> Outcome<String> { |
| 223 | if keys.is_empty() { |
| 224 | return Err(err!("A citation needs at least one key."; Input, Invalid, Missing)); |
| 225 | } |
| 226 | let suffixes = self.cited_suffixes(); |
| 227 | let mut labels: Vec<String> = Vec::with_capacity(keys.len()); |
| 228 | for &k in keys { |
| 229 | let entry = res!(self.entry(k).ok_or_else(|| |
| 230 | err!("Citation key {:?} is not in the bibliography.", k; Input, Missing))); |
| 231 | labels.push(cite_label(entry, suffixes.get(k).map(|s| s.as_str()).unwrap_or(""))); |
| 232 | } |
| 233 | Ok(fmt!("({})", labels.join("; "))) |
| 234 | } |
| 235 | |
| 236 | /// The in-text label for a single key without the surrounding parentheses, e.g. `Scott 1976`. |
| 237 | pub fn label(&self, key: &str) -> Outcome<String> { |
| 238 | let entry = res!(self.entry(key).ok_or_else(|| |
| 239 | err!("Citation key {:?} is not in the bibliography.", key; Input, Missing))); |
| 240 | let suffixes = self.cited_suffixes(); |
| 241 | Ok(cite_label(entry, suffixes.get(key).map(|s| s.as_str()).unwrap_or(""))) |
| 242 | } |
| 243 | |
| 244 | /// The year-disambiguation suffixes across the marked-cited works, keyed by entry key, so an in-text |
| 245 | /// citation reads the same `2019a`/`2019b` its reference does. Computed over the sorted cited set, which |
| 246 | /// is complete by the time citations are formatted (the reader marks every key before the walk sets them). |
| 247 | fn cited_suffixes(&self) -> BTreeMap<String, String> { |
| 248 | let mut refs: Vec<&Entry> = self.cited.iter() |
| 249 | .filter_map(|k| self.entry(k)) |
| 250 | .collect(); |
| 251 | sort_entries(&mut refs); |
| 252 | year_suffixes(&refs) |
| 253 | } |
| 254 | |
| 255 | /// The sorted, Chicago-styled reference list for the cited keys. If nothing has been marked |
| 256 | /// cited, the list is empty -- Typst prints only cited works. |
| 257 | pub fn reference_list(&self) -> Vec<Reference> { |
| 258 | let mut refs: Vec<&Entry> = self.cited.iter() |
| 259 | .filter_map(|k| self.entry(k)) |
| 260 | .collect(); |
| 261 | sort_entries(&mut refs); |
| 262 | let suffixes = year_suffixes(&refs); |
| 263 | refs.iter() |
| 264 | .map(|e| format_reference(e, suffixes.get(e.key()).map(|s| s.as_str()).unwrap_or(""))) |
| 265 | .collect() |
| 266 | } |
| 267 | |
| 268 | /// Every entry in the file, sorted and formatted. For tests and for a "print all" mode. |
| 269 | pub fn all_references(&self) -> Vec<Reference> { |
| 270 | let mut refs: Vec<&Entry> = self.entries.iter().collect(); |
| 271 | sort_entries(&mut refs); |
| 272 | let suffixes = year_suffixes(&refs); |
| 273 | refs.iter() |
| 274 | .map(|e| format_reference(e, suffixes.get(e.key()).map(|s| s.as_str()).unwrap_or(""))) |
| 275 | .collect() |
| 276 | } |
| 277 | } |
| 278 | |
| 279 | // --------------------------------------------------------------------------------------------- |
| 280 | // Parsing. |
| 281 | // --------------------------------------------------------------------------------------------- |
| 282 | |
| 283 | /// Parses one `@type{ key, field = value, ... }` starting at `chars[start]` (the `@`). Returns the |
| 284 | /// entry (or `None` for a `@string`/`@preamble`/`@comment` block) and the index just past its close. |
| 285 | fn parse_entry(chars: &[char], start: usize) -> Outcome<(Option<Entry>, usize)> { |
| 286 | let n = chars.len(); |
| 287 | let mut i = start + 1; // past '@' |
| 288 | // Read the type word. |
| 289 | let mut kind_word = String::new(); |
| 290 | while i < n && (chars[i].is_ascii_alphabetic()) { |
| 291 | kind_word.push(chars[i]); |
| 292 | i += 1; |
| 293 | } |
| 294 | // Skip to the opening brace or paren. |
| 295 | while i < n && chars[i] != '{' && chars[i] != '(' { |
| 296 | if chars[i] == '@' { |
| 297 | // A malformed entry with no body; bail without consuming the next '@'. |
| 298 | return Ok((None, i)); |
| 299 | } |
| 300 | i += 1; |
| 301 | } |
| 302 | if i >= n { |
| 303 | return Ok((None, n)); |
| 304 | } |
| 305 | let open = chars[i]; |
| 306 | let close = if open == '{' { '}' } else { ')' }; |
| 307 | i += 1; |
| 308 | |
| 309 | let lower = kind_word.to_ascii_lowercase(); |
| 310 | if lower == "string" || lower == "preamble" || lower == "comment" { |
| 311 | // Skip the balanced body. |
| 312 | let end = skip_balanced(chars, i, open, close); |
| 313 | return Ok((None, end)); |
| 314 | } |
| 315 | |
| 316 | // Read the citation key up to the first comma. |
| 317 | let mut key = String::new(); |
| 318 | while i < n && chars[i] != ',' && chars[i] != close { |
| 319 | if !chars[i].is_whitespace() { |
| 320 | key.push(chars[i]); |
| 321 | } |
| 322 | i += 1; |
| 323 | } |
| 324 | let mut fields: Vec<(String, String)> = Vec::new(); |
| 325 | // Read fields. |
| 326 | loop { |
| 327 | // Skip separators and whitespace. |
| 328 | while i < n && (chars[i] == ',' || chars[i].is_whitespace()) { |
| 329 | i += 1; |
| 330 | } |
| 331 | if i >= n || chars[i] == close { |
| 332 | if i < n { |
| 333 | i += 1; // consume close |
| 334 | } |
| 335 | break; |
| 336 | } |
| 337 | // Field name. |
| 338 | let mut name = String::new(); |
| 339 | while i < n && chars[i] != '=' && chars[i] != close && !chars[i].is_whitespace() { |
| 340 | name.push(chars[i]); |
| 341 | i += 1; |
| 342 | } |
| 343 | // Skip whitespace to '='. |
| 344 | while i < n && chars[i].is_whitespace() { |
| 345 | i += 1; |
| 346 | } |
| 347 | if i >= n || chars[i] != '=' { |
| 348 | // No value; a trailing key or malformed field. Stop at close if present. |
| 349 | if i < n && chars[i] == close { |
| 350 | i += 1; |
| 351 | } |
| 352 | break; |
| 353 | } |
| 354 | i += 1; // past '=' |
| 355 | while i < n && chars[i].is_whitespace() { |
| 356 | i += 1; |
| 357 | } |
| 358 | let (raw, next) = read_value(chars, i); |
| 359 | i = next; |
| 360 | if !name.is_empty() { |
| 361 | // `author` and `editor` are read by `parse_names`, which needs the BibTeX brace grouping intact: |
| 362 | // a corporate author is written `{{Name}}`, so after `read_value` strips the outer delimiter its |
| 363 | // value is one whole `{...}` group, and that surviving group is the only signal that the name is an |
| 364 | // organisation and must not be surname-inverted. `decode_value` collapses grouping braces, so |
| 365 | // decoding here would erase it and leave "International Energy Agency" to invert to "Agency, |
| 366 | // International Energy". These two fields are therefore stored raw; `parse_names` decodes each name |
| 367 | // part itself. Every other field is decoded now, as before. |
| 368 | let value = if name.eq_ignore_ascii_case("author") || name.eq_ignore_ascii_case("editor") { |
| 369 | raw |
| 370 | } else { |
| 371 | decode_value(&raw) |
| 372 | }; |
| 373 | fields.push((name.clone(), value)); |
| 374 | } |
| 375 | } |
| 376 | |
| 377 | let entry = Entry { |
| 378 | key, |
| 379 | kind: EntryKind::from_type(&lower), |
| 380 | fields, |
| 381 | }; |
| 382 | Ok((Some(entry), i)) |
| 383 | } |
| 384 | |
| 385 | /// Reads a field value: a `{...}` group, a `"..."` string, or a bare token (number or macro name). |
| 386 | /// Returns the raw inner text (braces or quotes stripped at the outermost level only) and the next |
| 387 | /// index. |
| 388 | fn read_value(chars: &[char], start: usize) -> (String, usize) { |
| 389 | let n = chars.len(); |
| 390 | let mut i = start; |
| 391 | if i >= n { |
| 392 | return (String::new(), i); |
| 393 | } |
| 394 | match chars[i] { |
| 395 | '{' => { |
| 396 | let mut depth = 0i32; |
| 397 | let mut out = String::new(); |
| 398 | while i < n { |
| 399 | let c = chars[i]; |
| 400 | if c == '{' { |
| 401 | depth += 1; |
| 402 | if depth > 1 { |
| 403 | out.push(c); |
| 404 | } |
| 405 | i += 1; |
| 406 | } else if c == '}' { |
| 407 | depth -= 1; |
| 408 | if depth == 0 { |
| 409 | i += 1; |
| 410 | break; |
| 411 | } |
| 412 | out.push(c); |
| 413 | i += 1; |
| 414 | } else { |
| 415 | out.push(c); |
| 416 | i += 1; |
| 417 | } |
| 418 | } |
| 419 | (out, i) |
| 420 | } |
| 421 | '"' => { |
| 422 | // Quoted value; braces inside protect nested quotes. |
| 423 | let mut depth = 0i32; |
| 424 | let mut out = String::new(); |
| 425 | i += 1; // past opening quote |
| 426 | while i < n { |
| 427 | let c = chars[i]; |
| 428 | if c == '{' { |
| 429 | depth += 1; |
| 430 | out.push(c); |
| 431 | i += 1; |
| 432 | } else if c == '}' { |
| 433 | depth -= 1; |
| 434 | out.push(c); |
| 435 | i += 1; |
| 436 | } else if c == '"' && depth == 0 { |
| 437 | i += 1; |
| 438 | break; |
| 439 | } else { |
| 440 | out.push(c); |
| 441 | i += 1; |
| 442 | } |
| 443 | } |
| 444 | (out, i) |
| 445 | } |
| 446 | _ => { |
| 447 | // Bare token to the next comma or closing brace/paren. |
| 448 | let mut out = String::new(); |
| 449 | while i < n && chars[i] != ',' && chars[i] != '}' && chars[i] != ')' { |
| 450 | out.push(chars[i]); |
| 451 | i += 1; |
| 452 | } |
| 453 | (out.trim().to_string(), i) |
| 454 | } |
| 455 | } |
| 456 | } |
| 457 | |
| 458 | /// Returns the index just past the balanced group opened at `i` (already inside the opener). |
| 459 | fn skip_balanced(chars: &[char], start: usize, open: char, close: char) -> usize { |
| 460 | let n = chars.len(); |
| 461 | let mut i = start; |
| 462 | let mut depth = 1i32; |
| 463 | while i < n && depth > 0 { |
| 464 | if chars[i] == open { |
| 465 | depth += 1; |
| 466 | } else if chars[i] == close { |
| 467 | depth -= 1; |
| 468 | } |
| 469 | i += 1; |
| 470 | } |
| 471 | i |
| 472 | } |
| 473 | |
| 474 | // --------------------------------------------------------------------------------------------- |
| 475 | // Value decoding: LaTeX accents, escapes, dashes and quotes to Unicode. |
| 476 | // --------------------------------------------------------------------------------------------- |
| 477 | |
| 478 | /// Decodes a raw BibTeX value to display text: resolves accent commands (`{\'e}` -> e-acute), the |
| 479 | /// literal escapes `\&`/`\%`/`\$` and the special letters (`\o`, `\L`, ...), collapses grouping |
| 480 | /// braces, and turns `--` into an en dash and TeX quotes into curly quotes. |
| 481 | fn decode_value(raw: &str) -> String { |
| 482 | let chars: Vec<char> = raw.chars().collect(); |
| 483 | let n = chars.len(); |
| 484 | let mut out = String::new(); |
| 485 | let mut i = 0usize; |
| 486 | while i < n { |
| 487 | let c = chars[i]; |
| 488 | match c { |
| 489 | '\\' => { |
| 490 | let (text, next) = decode_control(&chars, i); |
| 491 | out.push_str(&text); |
| 492 | i = next; |
| 493 | } |
| 494 | '{' | '}' => { |
| 495 | // Grouping brace: drop it, keep the contents. |
| 496 | i += 1; |
| 497 | } |
| 498 | '-' if i + 1 < n && chars[i + 1] == '-' => { |
| 499 | if i + 2 < n && chars[i + 2] == '-' { |
| 500 | out.push('\u{2014}'); // em dash |
| 501 | i += 3; |
| 502 | } else { |
| 503 | out.push('\u{2013}'); // en dash |
| 504 | i += 2; |
| 505 | } |
| 506 | } |
| 507 | '`' if i + 1 < n && chars[i + 1] == '`' => { |
| 508 | out.push('\u{201C}'); // opening double quote |
| 509 | i += 2; |
| 510 | } |
| 511 | '`' => { |
| 512 | out.push('\u{2018}'); // opening single quote |
| 513 | i += 1; |
| 514 | } |
| 515 | '\'' if i + 1 < n && chars[i + 1] == '\'' => { |
| 516 | out.push('\u{201D}'); // closing double quote |
| 517 | i += 2; |
| 518 | } |
| 519 | '~' => { |
| 520 | out.push('\u{00A0}'); // non-breaking space |
| 521 | i += 1; |
| 522 | } |
| 523 | _ => { |
| 524 | out.push(c); |
| 525 | i += 1; |
| 526 | } |
| 527 | } |
| 528 | } |
| 529 | out |
| 530 | } |
| 531 | |
| 532 | /// Decodes a control sequence beginning at `chars[i]` (a backslash). Returns the replacement text |
| 533 | /// and the index past what it consumed. |
| 534 | fn decode_control(chars: &[char], i: usize) -> (String, usize) { |
| 535 | let n = chars.len(); |
| 536 | let mut j = i + 1; // past backslash |
| 537 | if j >= n { |
| 538 | return (String::new(), j); |
| 539 | } |
| 540 | let c = chars[j]; |
| 541 | // Single-character escapes and accent markers. |
| 542 | if matches!(c, '&' | '%' | '$' | '#' | '_' | '{' | '}') { |
| 543 | return (c.to_string(), j + 1); |
| 544 | } |
| 545 | if matches!(c, '\'' | '`' | '"' | '^' | '~' | '=' | '.') { |
| 546 | // Accent taking the following letter, which may be `{x}` or bare `x`. |
| 547 | j += 1; |
| 548 | let (letter, next) = read_accent_arg(chars, j); |
| 549 | if let Some(a) = apply_accent(c, letter) { |
| 550 | return (a.to_string(), next); |
| 551 | } |
| 552 | return (letter.map(|l| l.to_string()).unwrap_or_default(), next); |
| 553 | } |
| 554 | // Alphabetic control word: `\c{c}`, `\v{s}`, `\L`, `\o`, `\ss`, ... |
| 555 | let mut word = String::new(); |
| 556 | while j < n && chars[j].is_ascii_alphabetic() { |
| 557 | word.push(chars[j]); |
| 558 | j += 1; |
| 559 | } |
| 560 | // Accent commands spelled as letters take an argument. |
| 561 | if matches!(word.as_str(), "c" | "v" | "u" | "H" | "r" | "k") { |
| 562 | let (letter, next) = read_accent_arg(chars, j); |
| 563 | if let Some(a) = apply_letter_accent(&word, letter) { |
| 564 | return (a.to_string(), next); |
| 565 | } |
| 566 | return (letter.map(|l| l.to_string()).unwrap_or_default(), next); |
| 567 | } |
| 568 | if let Some(s) = special_letter(&word) { |
| 569 | return (s.to_string(), j); |
| 570 | } |
| 571 | // Unknown command: drop the backslash, keep the word. |
| 572 | (word, j) |
| 573 | } |
| 574 | |
| 575 | /// Reads the argument of an accent: `{x}`, a bare letter, or nothing. Skips a leading space that |
| 576 | /// separates a control word from its letter. |
| 577 | fn read_accent_arg(chars: &[char], start: usize) -> (Option<char>, usize) { |
| 578 | let n = chars.len(); |
| 579 | let mut i = start; |
| 580 | while i < n && chars[i] == ' ' { |
| 581 | i += 1; |
| 582 | } |
| 583 | if i >= n { |
| 584 | return (None, i); |
| 585 | } |
| 586 | if chars[i] == '{' { |
| 587 | i += 1; |
| 588 | let letter = if i < n && chars[i] != '}' { Some(chars[i]) } else { None }; |
| 589 | if letter.is_some() { |
| 590 | i += 1; |
| 591 | } |
| 592 | if i < n && chars[i] == '}' { |
| 593 | i += 1; |
| 594 | } |
| 595 | (letter, i) |
| 596 | } else { |
| 597 | let letter = Some(chars[i]); |
| 598 | (letter, i + 1) |
| 599 | } |
| 600 | } |
| 601 | |
| 602 | fn apply_accent(mark: char, letter: Option<char>) -> Option<char> { |
| 603 | let l = letter?; |
| 604 | let r = match (mark, l) { |
| 605 | ('\'', 'a') => 'á', ('\'', 'e') => 'é', ('\'', 'i') => 'í', ('\'', 'o') => 'ó', |
| 606 | ('\'', 'u') => 'ú', ('\'', 'y') => 'ý', ('\'', 'n') => 'ń', ('\'', 'c') => 'ć', |
| 607 | ('\'', 's') => 'ś', ('\'', 'z') => 'ź', ('\'', 'A') => 'Á', ('\'', 'E') => 'É', |
| 608 | ('\'', 'I') => 'Í', ('\'', 'O') => 'Ó', ('\'', 'U') => 'Ú', |
| 609 | ('`', 'a') => 'à', ('`', 'e') => 'è', ('`', 'i') => 'ì', ('`', 'o') => 'ò', |
| 610 | ('`', 'u') => 'ù', ('`', 'A') => 'À', ('`', 'E') => 'È', ('`', 'O') => 'Ò', |
| 611 | ('"', 'a') => 'ä', ('"', 'e') => 'ë', ('"', 'i') => 'ï', ('"', 'o') => 'ö', |
| 612 | ('"', 'u') => 'ü', ('"', 'y') => 'ÿ', ('"', 'A') => 'Ä', ('"', 'O') => 'Ö', |
| 613 | ('"', 'U') => 'Ü', |
| 614 | ('^', 'a') => 'â', ('^', 'e') => 'ê', ('^', 'i') => 'î', ('^', 'o') => 'ô', |
| 615 | ('^', 'u') => 'û', ('^', 'A') => 'Â', ('^', 'O') => 'Ô', |
| 616 | ('~', 'a') => 'ã', ('~', 'n') => 'ñ', ('~', 'o') => 'õ', ('~', 'A') => 'Ã', |
| 617 | ('~', 'N') => 'Ñ', ('~', 'O') => 'Õ', |
| 618 | ('=', 'a') => 'ā', ('=', 'e') => 'ē', ('=', 'i') => 'ī', ('=', 'o') => 'ō', |
| 619 | ('=', 'u') => 'ū', |
| 620 | ('.', 'z') => 'ż', ('.', 'e') => 'ė', |
| 621 | _ => return None, |
| 622 | }; |
| 623 | Some(r) |
| 624 | } |
| 625 | |
| 626 | fn apply_letter_accent(cmd: &str, letter: Option<char>) -> Option<char> { |
| 627 | let l = letter?; |
| 628 | let r = match (cmd, l) { |
| 629 | ("c", 'c') => 'ç', ("c", 'C') => 'Ç', ("c", 's') => 'ş', ("c", 'S') => 'Ş', |
| 630 | ("v", 's') => 'š', ("v", 'S') => 'Š', ("v", 'c') => 'č', ("v", 'C') => 'Č', |
| 631 | ("v", 'z') => 'ž', ("v", 'Z') => 'Ž', ("v", 'r') => 'ř', ("v", 'e') => 'ě', |
| 632 | ("v", 'n') => 'ň', |
| 633 | ("u", 'a') => 'ă', ("u", 'g') => 'ğ', ("u", 'G') => 'Ğ', |
| 634 | ("H", 'o') => 'ő', ("H", 'u') => 'ű', |
| 635 | ("r", 'a') => 'å', ("r", 'A') => 'Å', ("r", 'u') => 'ů', |
| 636 | ("k", 'a') => 'ą', ("k", 'e') => 'ę', |
| 637 | _ => return None, |
| 638 | }; |
| 639 | Some(r) |
| 640 | } |
| 641 | |
| 642 | fn special_letter(word: &str) -> Option<char> { |
| 643 | let r = match word { |
| 644 | "L" => 'Ł', "l" => 'ł', "o" => 'ø', "O" => 'Ø', |
| 645 | "ss" => 'ß', "ae" => 'æ', "AE" => 'Æ', "oe" => 'œ', "OE" => 'Œ', |
| 646 | "aa" => 'å', "AA" => 'Å', "i" => 'ı', "j" => 'ȷ', "dh" => 'ð', "DH" => 'Ð', |
| 647 | "th" => 'þ', "TH" => 'Þ', |
| 648 | _ => return None, |
| 649 | }; |
| 650 | Some(r) |
| 651 | } |
| 652 | |
| 653 | // --------------------------------------------------------------------------------------------- |
| 654 | // Name parsing. |
| 655 | // --------------------------------------------------------------------------------------------- |
| 656 | |
| 657 | /// Parses a BibTeX author/editor field into names. Names are separated by ` and `; a name wrapped |
| 658 | /// in an extra brace group (`{{...}}` in the field) is corporate and kept whole. A personal name is |
| 659 | /// either `Family, Given` (comma form) or `Given ... Family` (natural form), the surname being the |
| 660 | /// text after the comma or the final whitespace-separated token. |
| 661 | fn parse_names(field: &str) -> Vec<Name> { |
| 662 | let mut names: Vec<Name> = Vec::new(); |
| 663 | for part in split_top(field) { |
| 664 | let trimmed = part.trim(); |
| 665 | if trimmed.is_empty() { |
| 666 | continue; |
| 667 | } |
| 668 | // Corporate: the whole part is one brace group. |
| 669 | if is_braced_whole(trimmed) { |
| 670 | let inner = &trimmed[1..trimmed.len() - 1]; |
| 671 | names.push(Name { |
| 672 | family: decode_value(inner), |
| 673 | given: String::new(), |
| 674 | corporate: true, |
| 675 | }); |
| 676 | continue; |
| 677 | } |
| 678 | if let Some(comma) = trimmed.find(',') { |
| 679 | let family = decode_value(trimmed[..comma].trim()); |
| 680 | let given = decode_value(trimmed[comma + 1..].trim()); |
| 681 | names.push(Name { family, given, corporate: false }); |
| 682 | } else { |
| 683 | // Natural order: last token is the surname. |
| 684 | let toks: Vec<&str> = trimmed.split_whitespace().collect(); |
| 685 | if toks.len() <= 1 { |
| 686 | names.push(Name { |
| 687 | family: decode_value(trimmed), |
| 688 | given: String::new(), |
| 689 | corporate: false, |
| 690 | }); |
| 691 | } else { |
| 692 | let family = decode_value(toks[toks.len() - 1]); |
| 693 | let given = decode_value(&toks[..toks.len() - 1].join(" ")); |
| 694 | names.push(Name { family, given, corporate: false }); |
| 695 | } |
| 696 | } |
| 697 | } |
| 698 | names |
| 699 | } |
| 700 | |
| 701 | /// Splits an author field on top-level ` and ` (not inside braces). |
| 702 | fn split_top(field: &str) -> Vec<String> { |
| 703 | let chars: Vec<char> = field.chars().collect(); |
| 704 | let n = chars.len(); |
| 705 | let mut parts: Vec<String> = Vec::new(); |
| 706 | let mut cur = String::new(); |
| 707 | let mut depth = 0i32; |
| 708 | let mut i = 0usize; |
| 709 | while i < n { |
| 710 | let c = chars[i]; |
| 711 | if c == '{' { |
| 712 | depth += 1; |
| 713 | cur.push(c); |
| 714 | i += 1; |
| 715 | } else if c == '}' { |
| 716 | depth -= 1; |
| 717 | cur.push(c); |
| 718 | i += 1; |
| 719 | } else if depth == 0 |
| 720 | && c == ' ' |
| 721 | && i + 4 < n |
| 722 | && chars[i + 1] == 'a' |
| 723 | && chars[i + 2] == 'n' |
| 724 | && chars[i + 3] == 'd' |
| 725 | && chars[i + 4] == ' ' |
| 726 | { |
| 727 | parts.push(cur.clone()); |
| 728 | cur.clear(); |
| 729 | i += 5; |
| 730 | } else { |
| 731 | cur.push(c); |
| 732 | i += 1; |
| 733 | } |
| 734 | } |
| 735 | if !cur.trim().is_empty() { |
| 736 | parts.push(cur); |
| 737 | } |
| 738 | parts |
| 739 | } |
| 740 | |
| 741 | fn is_braced_whole(s: &str) -> bool { |
| 742 | let chars: Vec<char> = s.chars().collect(); |
| 743 | if chars.len() < 2 || chars[0] != '{' || chars[chars.len() - 1] != '}' { |
| 744 | return false; |
| 745 | } |
| 746 | let mut depth = 0i32; |
| 747 | for (k, &c) in chars.iter().enumerate() { |
| 748 | if c == '{' { |
| 749 | depth += 1; |
| 750 | } else if c == '}' { |
| 751 | depth -= 1; |
| 752 | if depth == 0 && k != chars.len() - 1 { |
| 753 | return false; // the first group closes early: not one whole group |
| 754 | } |
| 755 | } |
| 756 | } |
| 757 | depth == 0 |
| 758 | } |
| 759 | |
| 760 | // --------------------------------------------------------------------------------------------- |
| 761 | // In-text citation label. |
| 762 | // --------------------------------------------------------------------------------------------- |
| 763 | |
| 764 | /// The author part of an in-text label, e.g. `Scott`, `Kahneman and Tversky`, `Acemoglu et al.`. Three |
| 765 | /// or more names collapse to the first plus `et al.`; an authorless work falls back to its title. This |
| 766 | /// is the key two references share when they must be disambiguated by a year suffix. |
| 767 | fn cite_who(entry: &Entry) -> String { |
| 768 | let names = entry.credited(); |
| 769 | match names.len() { |
| 770 | 0 => entry.field("title").map(|t| chicago_title_case(t)).unwrap_or_default(), |
| 771 | 1 => names[0].family().to_string(), |
| 772 | 2 => fmt!("{} and {}", names[0].family(), names[1].family()), |
| 773 | _ => fmt!("{} et al.", names[0].family()), |
| 774 | } |
| 775 | } |
| 776 | |
| 777 | /// The author-year label for one entry, e.g. `Scott 1976`, `Kahneman and Tversky 1979`, |
| 778 | /// `Acemoglu et al. 2001`. `suffix` is the year-disambiguation letter (`a`, `b`, ...) when this entry |
| 779 | /// shares an author and a year with another cited work, and empty otherwise, so `Zuboff 2019` becomes |
| 780 | /// `Zuboff 2019a`. |
| 781 | fn cite_label(entry: &Entry, suffix: &str) -> String { |
| 782 | let year = entry.field("year").map(|y| year_display(y)).unwrap_or_else(|| "n.d.".to_string()); |
| 783 | let year = fmt!("{}{}", year, suffix); |
| 784 | let who = cite_who(entry); |
| 785 | if who.is_empty() { |
| 786 | year |
| 787 | } else { |
| 788 | fmt!("{} {}", who, year) |
| 789 | } |
| 790 | } |
| 791 | |
| 792 | /// The Chicago year-disambiguation suffixes for a reference list: when two or more cited works share an |
| 793 | /// author label and a year, each takes a letter -- `a`, `b`, `c` ... -- in the order it stands in the |
| 794 | /// sorted list, so an in-text `(Zuboff 2019a)` matches its reference. A lone author-year takes none. The |
| 795 | /// returned map is keyed by entry key; a key absent from it has no suffix. `sorted` must already be in |
| 796 | /// reference-list order, since the letters follow it. |
| 797 | fn year_suffixes(sorted: &[&Entry]) -> BTreeMap<String, String> { |
| 798 | // Group the entries by (author label, year), keeping each group's members in the sorted order they |
| 799 | // arrived in; a group of two or more is ambiguous and its members take letters. |
| 800 | let mut order: Vec<Vec<String>> = Vec::new(); |
| 801 | let mut seen: BTreeMap<String, usize> = BTreeMap::new(); |
| 802 | for e in sorted { |
| 803 | let year = e.field("year").map(|y| year_display(y)).unwrap_or_else(|| "n.d.".to_string()); |
| 804 | let gid = fmt!("{}\u{0}{}", cite_who(e), year); |
| 805 | let pos = match seen.get(&gid) { |
| 806 | Some(&p) => p, |
| 807 | None => { |
| 808 | let p = order.len(); |
| 809 | order.push(Vec::new()); |
| 810 | seen.insert(gid, p); |
| 811 | p |
| 812 | }, |
| 813 | }; |
| 814 | order[pos].push(e.key().to_string()); |
| 815 | } |
| 816 | let mut out: BTreeMap<String, String> = BTreeMap::new(); |
| 817 | for group in &order { |
| 818 | if group.len() > 1 { |
| 819 | for (n, key) in group.iter().enumerate() { |
| 820 | out.insert(key.clone(), suffix_letter(n)); |
| 821 | } |
| 822 | } |
| 823 | } |
| 824 | out |
| 825 | } |
| 826 | |
| 827 | /// The `n`-th disambiguation letter: `0 -> a`, `25 -> z`, `26 -> aa`, and so on. A bibliography never |
| 828 | /// runs past a handful, but the base-26 roll-over keeps a pathological run well-formed rather than wrapping. |
| 829 | fn suffix_letter(mut n: usize) -> String { |
| 830 | let mut s = String::new(); |
| 831 | loop { |
| 832 | let d = (n % 26) as u8; |
| 833 | s.insert(0, (b'a' + d) as char); |
| 834 | if n < 26 { |
| 835 | break; |
| 836 | } |
| 837 | n = n / 26 - 1; |
| 838 | } |
| 839 | s |
| 840 | } |
| 841 | |
| 842 | /// The visible year: the leading run of digits (drops a BibTeX `{2024}` note or a trailing letter). |
| 843 | fn year_display(y: &str) -> String { |
| 844 | let digits: String = y.chars().take_while(|c| c.is_ascii_digit()).collect(); |
| 845 | if digits.is_empty() { y.trim().to_string() } else { digits } |
| 846 | } |
| 847 | |
| 848 | // --------------------------------------------------------------------------------------------- |
| 849 | // Reference list formatting (Chicago author-date). |
| 850 | // --------------------------------------------------------------------------------------------- |
| 851 | |
| 852 | /// A small builder that accumulates runs, merging consecutive same-style text. |
| 853 | struct RunBuilder { |
| 854 | runs: Vec<RefRun>, |
| 855 | } |
| 856 | |
| 857 | impl RunBuilder { |
| 858 | fn new() -> Self { Self { runs: Vec::new() } } |
| 859 | |
| 860 | fn push(&mut self, text: &str, style: RefStyle) { |
| 861 | if text.is_empty() { |
| 862 | return; |
| 863 | } |
| 864 | if let Some(last) = self.runs.last_mut() { |
| 865 | if last.style == style { |
| 866 | last.text.push_str(text); |
| 867 | return; |
| 868 | } |
| 869 | } |
| 870 | self.runs.push(RefRun { text: text.to_string(), style }); |
| 871 | } |
| 872 | |
| 873 | fn normal(&mut self, text: &str) { self.push(text, RefStyle::Normal); } |
| 874 | fn italic(&mut self, text: &str) { self.push(text, RefStyle::Italic); } |
| 875 | |
| 876 | fn finish(mut self, key: &str) -> Reference { |
| 877 | smartquote_runs(&mut self.runs); |
| 878 | Reference { key: key.to_string(), runs: self.runs } |
| 879 | } |
| 880 | } |
| 881 | |
| 882 | /// Formats one entry as a Chicago author-date reference. `suffix` is the year-disambiguation letter |
| 883 | /// (`a`, `b`, ...) when this entry shares an author and a year with another in the list, empty otherwise. |
| 884 | fn format_reference(entry: &Entry, suffix: &str) -> Reference { |
| 885 | let mut b = RunBuilder::new(); |
| 886 | // Author. Year. |
| 887 | let author = author_list(&entry.credited(), entry.field("author").is_none()); |
| 888 | if !author.is_empty() { |
| 889 | b.normal(&author); |
| 890 | // A name ending in an initial already carries its full stop. |
| 891 | if author.ends_with('.') { |
| 892 | b.normal(" "); |
| 893 | } else { |
| 894 | b.normal(". "); |
| 895 | } |
| 896 | } |
| 897 | if let Some(y) = entry.field("year") { |
| 898 | b.normal(&fmt!("{}{}", year_display(y), suffix)); |
| 899 | b.normal(". "); |
| 900 | } |
| 901 | match entry.kind { |
| 902 | EntryKind::Article => format_article(&mut b, entry), |
| 903 | EntryKind::InCollection | EntryKind::InProceedings => format_incollection(&mut b, entry), |
| 904 | EntryKind::TechReport | EntryKind::Report => format_report(&mut b, entry), |
| 905 | EntryKind::Online => format_online(&mut b, entry), |
| 906 | _ => format_book(&mut b, entry), |
| 907 | } |
| 908 | b.finish(entry.key()) |
| 909 | } |
| 910 | |
| 911 | /// The author string for the list: first name inverted, the rest natural, `and` before the last, |
| 912 | /// with an `, ed.`/`, eds.` tag when the credited names are editors. |
| 913 | fn author_list(names: &[Name], is_editor: bool) -> String { |
| 914 | let body = match names.len() { |
| 915 | 0 => String::new(), |
| 916 | 1 => names[0].inverted(), |
| 917 | 2 => fmt!("{}, and {}", names[0].inverted(), names[1].natural()), |
| 918 | _ => { |
| 919 | let mut s = names[0].inverted(); |
| 920 | for name in &names[1..names.len() - 1] { |
| 921 | s.push_str(", "); |
| 922 | s.push_str(&name.natural()); |
| 923 | } |
| 924 | s.push_str(", and "); |
| 925 | s.push_str(&names[names.len() - 1].natural()); |
| 926 | s |
| 927 | } |
| 928 | }; |
| 929 | if is_editor && !body.is_empty() { |
| 930 | let tag = if names.len() > 1 { ", eds." } else { ", ed." }; |
| 931 | fmt!("{}{}", body, tag) |
| 932 | } else { |
| 933 | body |
| 934 | } |
| 935 | } |
| 936 | |
| 937 | fn format_book(b: &mut RunBuilder, entry: &Entry) { |
| 938 | if let Some(t) = entry.field("title") { |
| 939 | b.italic(&chicago_title_case(t)); |
| 940 | b.normal(". "); |
| 941 | } |
| 942 | // Chicago author-date omits the place for a book with a named publisher. |
| 943 | if let Some(p) = entry.field("publisher") { |
| 944 | b.normal(p); |
| 945 | b.normal("."); |
| 946 | } else if let Some(a) = entry.field("address") { |
| 947 | b.normal(a); |
| 948 | b.normal("."); |
| 949 | } |
| 950 | append_url(b, entry); |
| 951 | } |
| 952 | |
| 953 | fn format_article(b: &mut RunBuilder, entry: &Entry) { |
| 954 | if let Some(t) = entry.field("title") { |
| 955 | b.normal("\u{201C}"); |
| 956 | b.normal(&chicago_title_case(t)); |
| 957 | b.normal(".\u{201D} "); |
| 958 | } |
| 959 | if let Some(j) = entry.field("journal").or_else(|| entry.field("journaltitle")) { |
| 960 | b.italic(&chicago_title_case(j)); |
| 961 | match entry.field("volume") { |
| 962 | Some(v) => { |
| 963 | b.normal(" "); |
| 964 | b.normal(v); |
| 965 | if let Some(num) = entry.field("number") { |
| 966 | b.normal(&fmt!(" ({})", num)); |
| 967 | } |
| 968 | if let Some(p) = entry.field("pages") { |
| 969 | b.normal(&fmt!(": {}", compress_pages(p))); |
| 970 | } |
| 971 | b.normal("."); |
| 972 | } |
| 973 | // Typst sets a bare comma after the journal when there is no volume. |
| 974 | None => b.normal(",."), |
| 975 | } |
| 976 | } |
| 977 | append_url(b, entry); |
| 978 | } |
| 979 | |
| 980 | fn format_incollection(b: &mut RunBuilder, entry: &Entry) { |
| 981 | if let Some(t) = entry.field("title") { |
| 982 | b.normal("\u{201C}"); |
| 983 | b.normal(&chicago_title_case(t)); |
| 984 | b.normal(".\u{201D} "); |
| 985 | } |
| 986 | b.normal("In "); |
| 987 | if let Some(bt) = entry.field("booktitle") { |
| 988 | b.italic(&chicago_title_case(bt)); |
| 989 | } |
| 990 | if let Some(ed) = entry.field("editor") { |
| 991 | let eds = parse_names(ed); |
| 992 | b.normal(&fmt!(", edited by {}", natural_join(&eds))); |
| 993 | } |
| 994 | if let Some(p) = entry.field("pages") { |
| 995 | b.normal(&fmt!(", {}", compress_pages(p))); |
| 996 | } |
| 997 | b.normal(". "); |
| 998 | if let Some(p) = entry.field("publisher") { |
| 999 | b.normal(p); |
| 1000 | b.normal("."); |
| 1001 | } else if let Some(a) = entry.field("address") { |
| 1002 | b.normal(a); |
| 1003 | b.normal("."); |
| 1004 | } |
| 1005 | append_url(b, entry); |
| 1006 | } |
| 1007 | |
| 1008 | fn format_report(b: &mut RunBuilder, entry: &Entry) { |
| 1009 | if let Some(t) = entry.field("title") { |
| 1010 | b.italic(&chicago_title_case(t)); |
| 1011 | b.normal(". "); |
| 1012 | } |
| 1013 | let kind = entry.field("type").unwrap_or("Working Paper"); |
| 1014 | match entry.field("number") { |
| 1015 | Some(num) => b.normal(&fmt!("{} No. {}. ", kind, num)), |
| 1016 | None => { |
| 1017 | b.normal(kind); |
| 1018 | b.normal(". "); |
| 1019 | } |
| 1020 | } |
| 1021 | if let Some(inst) = entry.field("institution") { |
| 1022 | // The address stands in for the imprint when present; otherwise the institution. |
| 1023 | match entry.field("address") { |
| 1024 | Some(a) => { |
| 1025 | b.normal(a); |
| 1026 | b.normal("."); |
| 1027 | } |
| 1028 | None => { |
| 1029 | b.normal(inst); |
| 1030 | b.normal("."); |
| 1031 | } |
| 1032 | } |
| 1033 | } else if let Some(a) = entry.field("address") { |
| 1034 | b.normal(a); |
| 1035 | b.normal("."); |
| 1036 | } |
| 1037 | append_url(b, entry); |
| 1038 | } |
| 1039 | |
| 1040 | fn format_online(b: &mut RunBuilder, entry: &Entry) { |
| 1041 | if let Some(t) = entry.field("title") { |
| 1042 | b.italic(&chicago_title_case(t)); |
| 1043 | b.normal(". "); |
| 1044 | } |
| 1045 | if let Some(pubr) = entry.field("publisher").or_else(|| entry.field("organization")) { |
| 1046 | b.normal(pubr); |
| 1047 | b.normal(". "); |
| 1048 | } |
| 1049 | if let Some(u) = entry.field("url") { |
| 1050 | b.normal(u); |
| 1051 | b.normal("."); |
| 1052 | } |
| 1053 | } |
| 1054 | |
| 1055 | /// Appends a DOI (as an `https://doi.org/` URL) or a bare URL, whichever is present. |
| 1056 | fn append_url(b: &mut RunBuilder, entry: &Entry) { |
| 1057 | if let Some(doi) = entry.field("doi") { |
| 1058 | b.normal(&fmt!(" https://doi.org/{}.", doi.trim())); |
| 1059 | } else if let Some(u) = entry.field("url") { |
| 1060 | b.normal(&fmt!(" {}.", u.trim())); |
| 1061 | } |
| 1062 | } |
| 1063 | |
| 1064 | /// Joins names in natural order with commas and a final `and`. |
| 1065 | fn natural_join(names: &[Name]) -> String { |
| 1066 | match names.len() { |
| 1067 | 0 => String::new(), |
| 1068 | 1 => names[0].natural(), |
| 1069 | 2 => fmt!("{} and {}", names[0].natural(), names[1].natural()), |
| 1070 | _ => { |
| 1071 | let mut s = String::new(); |
| 1072 | for name in &names[..names.len() - 1] { |
| 1073 | s.push_str(&name.natural()); |
| 1074 | s.push_str(", "); |
| 1075 | } |
| 1076 | s.push_str("and "); |
| 1077 | s.push_str(&names[names.len() - 1].natural()); |
| 1078 | s |
| 1079 | } |
| 1080 | } |
| 1081 | } |
| 1082 | |
| 1083 | // --------------------------------------------------------------------------------------------- |
| 1084 | // Sorting. |
| 1085 | // --------------------------------------------------------------------------------------------- |
| 1086 | |
| 1087 | /// Sorts references Chicago-style. Works are ordered by the first author's surname; among works by |
| 1088 | /// the same first author, a work with fewer authors comes first (a solo work before coauthored |
| 1089 | /// ones), then by the coauthors' surnames, then by year, then by title. This author-count tie-break |
| 1090 | /// is Typst's behaviour and was verified against a render. |
| 1091 | fn sort_entries(entries: &mut [&Entry]) { |
| 1092 | entries.sort_by(|a, b| sort_key(a).cmp(&sort_key(b))); |
| 1093 | } |
| 1094 | |
| 1095 | fn sort_key(entry: &Entry) -> (String, usize, Vec<String>, i64, String) { |
| 1096 | let names = entry.credited(); |
| 1097 | let first = names.first().map(|n| fold_key(n.family())).unwrap_or_default(); |
| 1098 | let count = names.len(); |
| 1099 | let rest: Vec<String> = names.iter().skip(1).map(|n| fold_key(n.family())).collect(); |
| 1100 | let year = entry.year().unwrap_or(i64::MAX); |
| 1101 | let title = entry.field("title").map(|t| fold_key(t)).unwrap_or_default(); |
| 1102 | (first, count, rest, year, title) |
| 1103 | } |
| 1104 | |
| 1105 | /// A case- and accent-folded sort key for a surname or title. |
| 1106 | fn fold_key(s: &str) -> String { |
| 1107 | s.chars() |
| 1108 | .filter(|c| c.is_alphanumeric() || *c == ' ') |
| 1109 | .flat_map(|c| c.to_lowercase()) |
| 1110 | .collect() |
| 1111 | } |
| 1112 | |
| 1113 | // --------------------------------------------------------------------------------------------- |
| 1114 | // Chicago page-range abbreviation. |
| 1115 | // --------------------------------------------------------------------------------------------- |
| 1116 | |
| 1117 | /// Abbreviates a page range by the Chicago rule and sets the separator as an en dash. `263--291` |
| 1118 | /// becomes `263-91`, `1369--1401` becomes `1369-401`, and `488--500` stays `488-500`. A range whose |
| 1119 | /// ends differ in length, or a non-numeric range, keeps both ends whole. |
| 1120 | fn compress_pages(pages: &str) -> String { |
| 1121 | let norm = pages.replace("--", "-").replace('\u{2013}', "-"); |
| 1122 | let bits: Vec<&str> = norm.splitn(2, '-').collect(); |
| 1123 | if bits.len() != 2 { |
| 1124 | return norm.trim().to_string(); |
| 1125 | } |
| 1126 | let a = bits[0].trim(); |
| 1127 | let z = bits[1].trim(); |
| 1128 | let en = '\u{2013}'; |
| 1129 | let (an, zn) = (a.parse::<i64>(), z.parse::<i64>()); |
| 1130 | match (an, zn) { |
| 1131 | (Ok(av), Ok(_)) if a.len() == z.len() => { |
| 1132 | if av < 100 || av % 100 == 0 { |
| 1133 | fmt!("{}{}{}", a, en, z) |
| 1134 | } else { |
| 1135 | let ac: Vec<char> = a.chars().collect(); |
| 1136 | let zc: Vec<char> = z.chars().collect(); |
| 1137 | let mut common = 0usize; |
| 1138 | while common < ac.len() && ac[common] == zc[common] { |
| 1139 | common += 1; |
| 1140 | } |
| 1141 | let min_keep = if av % 100 <= 9 { 1 } else { 2 }; |
| 1142 | let keep = std::cmp::max(zc.len() - common, min_keep); |
| 1143 | let tail: String = zc[zc.len() - keep..].iter().collect(); |
| 1144 | fmt!("{}{}{}", a, en, tail) |
| 1145 | } |
| 1146 | } |
| 1147 | _ => fmt!("{}{}{}", a, en, z), |
| 1148 | } |
| 1149 | } |
| 1150 | |
| 1151 | // --------------------------------------------------------------------------------------------- |
| 1152 | // Chicago headline-style title casing. |
| 1153 | // --------------------------------------------------------------------------------------------- |
| 1154 | |
| 1155 | /// Downcases the minor words of an already-capitalised title to Typst's Chicago headline style, |
| 1156 | /// e.g. `Decision Under Risk` becomes `Decision under Risk`. The first word, and the first word |
| 1157 | /// after a colon, stay capitalised; a minor word anywhere else is downcased, including in final |
| 1158 | /// position (`Bringing the State Back In` becomes `... Back in`). Other words keep their source |
| 1159 | /// capitalisation, so a deliberately capitalised acronym in the source is preserved. |
| 1160 | fn chicago_title_case(title: &str) -> String { |
| 1161 | let words: Vec<&str> = title.split(' ').collect(); |
| 1162 | let mut out: Vec<String> = Vec::with_capacity(words.len()); |
| 1163 | let mut start_of_clause = true; // first word, or first after a colon |
| 1164 | for (k, w) in words.iter().enumerate() { |
| 1165 | let lowered = w.to_lowercase(); |
| 1166 | let bare: String = lowered.chars().filter(|c| c.is_alphabetic()).collect(); |
| 1167 | let is_minor = MINOR_WORDS.contains(&bare.as_str()); |
| 1168 | // Typst downcases a minor word wherever it falls, save the start of a clause -- so even |
| 1169 | // a trailing "In" becomes "in" ("Bringing the State Back in"). |
| 1170 | let downcase = is_minor && !start_of_clause && k != 0; |
| 1171 | if downcase { |
| 1172 | out.push(lowered); |
| 1173 | } else { |
| 1174 | out.push(w.to_string()); |
| 1175 | } |
| 1176 | // The next word starts a clause if this word ends with a colon. |
| 1177 | start_of_clause = w.ends_with(':'); |
| 1178 | } |
| 1179 | out.join(" ") |
| 1180 | } |
| 1181 | |
| 1182 | /// The articles, coordinating conjunctions and short prepositions Chicago sets lower case when they |
| 1183 | /// fall inside a title. |
| 1184 | const MINOR_WORDS: &[&str] = &[ |
| 1185 | "a", "an", "the", |
| 1186 | "and", "but", "or", "nor", "for", "so", "yet", |
| 1187 | "as", "at", "by", "in", "of", "off", "on", "per", "to", "up", "via", |
| 1188 | "from", "into", "like", "near", "onto", "over", "than", "that", "till", |
| 1189 | "unto", "upon", "with", "about", "above", "after", "among", "under", |
| 1190 | ]; |
| 1191 | |
| 1192 | // --------------------------------------------------------------------------------------------- |
| 1193 | // Smart quotes. |
| 1194 | // --------------------------------------------------------------------------------------------- |
| 1195 | |
| 1196 | /// Turns straight apostrophes and straight double quotes in the runs into curly quotes, matching |
| 1197 | /// Typst's smart-quote pass. TeX quote pairs were already converted in [`decode_value`]. |
| 1198 | fn smartquote_runs(runs: &mut [RefRun]) { |
| 1199 | for run in runs.iter_mut() { |
| 1200 | run.text = smartquote(&run.text); |
| 1201 | } |
| 1202 | } |
| 1203 | |
| 1204 | fn smartquote(s: &str) -> String { |
| 1205 | let chars: Vec<char> = s.chars().collect(); |
| 1206 | let n = chars.len(); |
| 1207 | let mut out = String::with_capacity(n); |
| 1208 | let mut dq_open = false; |
| 1209 | for (i, &c) in chars.iter().enumerate() { |
| 1210 | match c { |
| 1211 | '\'' => { |
| 1212 | let prev_alnum = i > 0 && chars[i - 1].is_alphanumeric(); |
| 1213 | if prev_alnum { |
| 1214 | out.push('\u{2019}'); // apostrophe / closing single |
| 1215 | } else { |
| 1216 | out.push('\u{2018}'); |
| 1217 | } |
| 1218 | } |
| 1219 | '"' => { |
| 1220 | if dq_open { |
| 1221 | out.push('\u{201D}'); |
| 1222 | dq_open = false; |
| 1223 | } else { |
| 1224 | out.push('\u{201C}'); |
| 1225 | dq_open = true; |
| 1226 | } |
| 1227 | } |
| 1228 | _ => out.push(c), |
| 1229 | } |
| 1230 | } |
| 1231 | out |
| 1232 | } |
| 1233 | |
| 1234 | #[cfg(test)] |
| 1235 | mod tests { |
| 1236 | use super::*; |
| 1237 | |
| 1238 | const SAMPLE: &str = r#" |
| 1239 | % A comment line that must be skipped. |
| 1240 | @book{scott1976moral, |
| 1241 | author = {Scott, James C.}, |
| 1242 | title = {The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia}, |
| 1243 | year = {1976}, |
| 1244 | publisher = {Yale University Press}, |
| 1245 | address = {New Haven} |
| 1246 | } |
| 1247 | |
| 1248 | @article{kahneman1979prospect, |
| 1249 | author = {Kahneman, Daniel and Tversky, Amos}, |
| 1250 | title = {Prospect Theory: An Analysis of Decision Under Risk}, |
| 1251 | journal = {Econometrica}, |
| 1252 | volume = {47}, |
| 1253 | number = {2}, |
| 1254 | pages = {263--291}, |
| 1255 | year = {1979}, |
| 1256 | doi = {10.2307/1914185} |
| 1257 | } |
| 1258 | |
| 1259 | @article{acemoglu2001colonial, |
| 1260 | author = {Acemoglu, Daron and Johnson, Simon and Robinson, James A.}, |
| 1261 | title = {The Colonial Origins of Comparative Development: An Empirical Investigation}, |
| 1262 | journal = {American Economic Review}, |
| 1263 | volume = {91}, |
| 1264 | number = {5}, |
| 1265 | pages = {1369--1401}, |
| 1266 | year = {2001}, |
| 1267 | doi = {10.1257/aer.91.5.1369} |
| 1268 | } |
| 1269 | |
| 1270 | @techreport{acemoglu2024simple, |
| 1271 | author = {Acemoglu, Daron}, |
| 1272 | title = {The Simple Macroeconomics of {AI}}, |
| 1273 | institution = {National Bureau of Economic Research}, |
| 1274 | type = {Working Paper}, |
| 1275 | number = {32487}, |
| 1276 | year = {2024}, |
| 1277 | address = {Cambridge, MA} |
| 1278 | } |
| 1279 | |
| 1280 | @book{worldbank2024, |
| 1281 | author = {World Bank}, |
| 1282 | title = {Education Finance Watch 2024}, |
| 1283 | year = {2024}, |
| 1284 | institution = {World Bank Group} |
| 1285 | } |
| 1286 | |
| 1287 | @incollection{tilly1985war, |
| 1288 | author = {Tilly, Charles}, |
| 1289 | title = {War Making and State Making as Organized Crime}, |
| 1290 | booktitle = {Bringing the State Back In}, |
| 1291 | editor = {Evans, Peter B. and Rueschemeyer, Dietrich and Skocpol, Theda}, |
| 1292 | publisher = {Cambridge University Press}, |
| 1293 | address = {Cambridge}, |
| 1294 | year = {1985}, |
| 1295 | pages = {169--191} |
| 1296 | } |
| 1297 | |
| 1298 | @book{kornai1992socialist, |
| 1299 | author = {Kornai, J{\'a}nos}, |
| 1300 | title = {The Socialist System}, |
| 1301 | year = {1992}, |
| 1302 | publisher = {Princeton University Press} |
| 1303 | } |
| 1304 | "#; |
| 1305 | |
| 1306 | fn bib() -> Bibliography { |
| 1307 | match Bibliography::parse(SAMPLE) { |
| 1308 | Ok(b) => b, |
| 1309 | Err(e) => panic!("parse failed: {}", e), |
| 1310 | } |
| 1311 | } |
| 1312 | |
| 1313 | #[test] |
| 1314 | fn parses_all_entries() { |
| 1315 | let b = bib(); |
| 1316 | assert_eq!(b.len(), 7); |
| 1317 | let scott = b.entry("scott1976moral").expect("scott present"); |
| 1318 | assert_eq!(scott.kind(), EntryKind::Book); |
| 1319 | assert_eq!(scott.field("year"), Some("1976")); |
| 1320 | assert_eq!(scott.field("publisher"), Some("Yale University Press")); |
| 1321 | } |
| 1322 | |
| 1323 | #[test] |
| 1324 | fn decodes_accent_and_brace() { |
| 1325 | let b = bib(); |
| 1326 | let k = b.entry("kornai1992socialist").expect("kornai present"); |
| 1327 | let names = k.credited(); |
| 1328 | assert_eq!(names[0].family(), "Kornai"); |
| 1329 | assert_eq!(names[0].given(), "János"); // {\'a} decoded |
| 1330 | // The braced {AI} keeps its letters, braces dropped. |
| 1331 | let tr = b.entry("acemoglu2024simple").expect("report present"); |
| 1332 | assert_eq!(tr.field("title"), Some("The Simple Macroeconomics of AI")); |
| 1333 | } |
| 1334 | |
| 1335 | #[test] |
| 1336 | fn parses_multiple_authors() { |
| 1337 | let b = bib(); |
| 1338 | let a = b.entry("acemoglu2001colonial").expect("present"); |
| 1339 | let names = a.credited(); |
| 1340 | assert_eq!(names.len(), 3); |
| 1341 | assert_eq!(names[1].family(), "Johnson"); |
| 1342 | assert_eq!(names[1].given(), "Simon"); |
| 1343 | } |
| 1344 | |
| 1345 | #[test] |
| 1346 | fn corporate_single_brace_is_inverted_as_personal() { |
| 1347 | // Typst reads `{World Bank}` as First=World, Last=Bank and inverts it in the list. |
| 1348 | let b = bib(); |
| 1349 | let refr = format_reference(b.entry("worldbank2024").expect("present"), ""); |
| 1350 | assert!(refr.plain().starts_with("Bank, World. 2024."), |
| 1351 | "got: {}", refr.plain()); |
| 1352 | } |
| 1353 | |
| 1354 | #[test] |
| 1355 | fn corporate_double_brace_is_not_inverted() { |
| 1356 | // A BibTeX corporate author is written `{{...}}`; after `read_value` strips the outer delimiter the |
| 1357 | // surviving brace group marks it an organisation, which Chicago prints verbatim -- never inverted to |
| 1358 | // "Agency, International Energy". Regression for the decode-before-parse bug that erased the group. |
| 1359 | let src = r#" |
| 1360 | @report{iea2024, |
| 1361 | author = {{International Energy Agency}}, |
| 1362 | title = {Electricity 2024}, |
| 1363 | year = {2024}, |
| 1364 | publisher = {IEA} |
| 1365 | } |
| 1366 | "#; |
| 1367 | let b = match Bibliography::parse(src) { Ok(b) => b, Err(e) => panic!("parse: {}", e) }; |
| 1368 | let r = format_reference(b.entry("iea2024").expect("present"), ""); |
| 1369 | assert!(r.plain().starts_with("International Energy Agency. 2024."), |
| 1370 | "corporate author must print verbatim; got: {}", r.plain()); |
| 1371 | let names = b.entry("iea2024").expect("present").credited(); |
| 1372 | assert_eq!(names.len(), 1); |
| 1373 | assert_eq!(names[0].family(), "International Energy Agency"); |
| 1374 | assert!(names[0].given().is_empty()); |
| 1375 | } |
| 1376 | |
| 1377 | #[test] |
| 1378 | fn same_author_same_year_takes_letter_suffix() { |
| 1379 | // Two cited works by one author in one year disambiguate as 2019a / 2019b, and an in-text citation |
| 1380 | // reads the same letter its reference does. A lone author-year (2020) takes no suffix. |
| 1381 | let src = r#" |
| 1382 | @book{zuboffA, |
| 1383 | author = {Zuboff, Shoshana}, |
| 1384 | title = {The Age of Surveillance Capitalism}, |
| 1385 | year = {2019}, |
| 1386 | publisher = {PublicAffairs} |
| 1387 | } |
| 1388 | @book{zuboffB, |
| 1389 | author = {Zuboff, Shoshana}, |
| 1390 | title = {Big Other}, |
| 1391 | year = {2019}, |
| 1392 | publisher = {Profile Books} |
| 1393 | } |
| 1394 | @book{zuboffC, |
| 1395 | author = {Zuboff, Shoshana}, |
| 1396 | title = {In the Age of the Smart Machine}, |
| 1397 | year = {1988}, |
| 1398 | publisher = {Basic Books} |
| 1399 | } |
| 1400 | "#; |
| 1401 | let mut b = match Bibliography::parse(src) { Ok(b) => b, Err(e) => panic!("parse: {}", e) }; |
| 1402 | let _ = b.cite(&["zuboffA", "zuboffB", "zuboffC"]).expect("cite"); |
| 1403 | let la = b.label("zuboffA").expect("label a"); |
| 1404 | let lb = b.label("zuboffB").expect("label b"); |
| 1405 | let lc = b.label("zuboffC").expect("label c"); |
| 1406 | // The 2019 pair each takes a distinct letter; the 1988 lone work takes none. |
| 1407 | assert!(la.ends_with("2019a") || la.ends_with("2019b"), "got {}", la); |
| 1408 | assert!(lb.ends_with("2019a") || lb.ends_with("2019b"), "got {}", lb); |
| 1409 | assert_ne!(la, lb); |
| 1410 | assert_eq!(lc, "Zuboff 1988"); |
| 1411 | // The reference list carries both suffixes, and each in-text year matches its own reference. |
| 1412 | let list = b.reference_list(); |
| 1413 | let all: String = list.iter().map(|r| r.plain()).collect::<Vec<_>>().join("\n"); |
| 1414 | assert!(all.contains("2019a."), "list missing 2019a: {}", all); |
| 1415 | assert!(all.contains("2019b."), "list missing 2019b: {}", all); |
| 1416 | assert!(!all.contains("1988a."), "lone year must not take a suffix: {}", all); |
| 1417 | } |
| 1418 | |
| 1419 | #[test] |
| 1420 | fn in_text_labels() { |
| 1421 | let b = bib(); |
| 1422 | assert_eq!(b.label("scott1976moral").expect("ok"), "Scott 1976"); |
| 1423 | assert_eq!(b.label("kahneman1979prospect").expect("ok"), "Kahneman and Tversky 1979"); |
| 1424 | assert_eq!(b.label("acemoglu2001colonial").expect("ok"), "Acemoglu et al. 2001"); |
| 1425 | } |
| 1426 | |
| 1427 | #[test] |
| 1428 | fn citation_single_and_multiple() { |
| 1429 | let mut b = bib(); |
| 1430 | assert_eq!(b.cite(&["scott1976moral"]).expect("ok"), "(Scott 1976)"); |
| 1431 | let multi = b.cite(&["scott1976moral", "kahneman1979prospect"]).expect("ok"); |
| 1432 | assert_eq!(multi, "(Scott 1976; Kahneman and Tversky 1979)"); |
| 1433 | } |
| 1434 | |
| 1435 | #[test] |
| 1436 | fn unknown_key_errors() { |
| 1437 | let b = bib(); |
| 1438 | assert!(b.format_citation(&["nope"]).is_err()); |
| 1439 | } |
| 1440 | |
| 1441 | #[test] |
| 1442 | fn book_reference_form() { |
| 1443 | let b = bib(); |
| 1444 | let r = format_reference(b.entry("scott1976moral").expect("present"), ""); |
| 1445 | assert_eq!( |
| 1446 | r.plain(), |
| 1447 | "Scott, James C. 1976. The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia. Yale University Press."); |
| 1448 | // The title is one italic run. |
| 1449 | assert!(r.runs.iter().any(|run| run.style == RefStyle::Italic |
| 1450 | && run.text.starts_with("The Moral Economy"))); |
| 1451 | } |
| 1452 | |
| 1453 | #[test] |
| 1454 | fn article_reference_form_with_page_compression() { |
| 1455 | let b = bib(); |
| 1456 | let r = format_reference(b.entry("kahneman1979prospect").expect("present"), ""); |
| 1457 | // Title down-cased "Under" -> "under"; pages 263--291 -> 263-91; DOI as a URL. |
| 1458 | assert_eq!( |
| 1459 | r.plain(), |
| 1460 | "Kahneman, Daniel, and Amos Tversky. 1979. \u{201C}Prospect Theory: An Analysis of Decision under Risk.\u{201D} Econometrica 47 (2): 263\u{2013}91. https://doi.org/10.2307/1914185."); |
| 1461 | } |
| 1462 | |
| 1463 | #[test] |
| 1464 | fn three_author_article_and_long_page_range() { |
| 1465 | let b = bib(); |
| 1466 | let r = format_reference(b.entry("acemoglu2001colonial").expect("present"), ""); |
| 1467 | assert_eq!( |
| 1468 | r.plain(), |
| 1469 | "Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. \u{201C}The Colonial Origins of Comparative Development: An Empirical Investigation.\u{201D} American Economic Review 91 (5): 1369\u{2013}401. https://doi.org/10.1257/aer.91.5.1369."); |
| 1470 | } |
| 1471 | |
| 1472 | #[test] |
| 1473 | fn report_reference_form() { |
| 1474 | let b = bib(); |
| 1475 | let r = format_reference(b.entry("acemoglu2024simple").expect("present"), ""); |
| 1476 | assert_eq!( |
| 1477 | r.plain(), |
| 1478 | "Acemoglu, Daron. 2024. The Simple Macroeconomics of AI. Working Paper No. 32487. Cambridge, MA."); |
| 1479 | } |
| 1480 | |
| 1481 | #[test] |
| 1482 | fn incollection_reference_form() { |
| 1483 | let b = bib(); |
| 1484 | let r = format_reference(b.entry("tilly1985war").expect("present"), ""); |
| 1485 | assert_eq!( |
| 1486 | r.plain(), |
| 1487 | "Tilly, Charles. 1985. \u{201C}War Making and State Making as Organized Crime.\u{201D} In Bringing the State Back in, edited by Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol, 169\u{2013}91. Cambridge University Press."); |
| 1488 | } |
| 1489 | |
| 1490 | #[test] |
| 1491 | fn page_compression_rule() { |
| 1492 | assert_eq!(compress_pages("263--291"), "263\u{2013}91"); |
| 1493 | assert_eq!(compress_pages("1369--1401"), "1369\u{2013}401"); |
| 1494 | assert_eq!(compress_pages("488--500"), "488\u{2013}500"); |
| 1495 | assert_eq!(compress_pages("855--857"), "855\u{2013}57"); |
| 1496 | assert_eq!(compress_pages("224--232"), "224\u{2013}32"); |
| 1497 | assert_eq!(compress_pages("97--112"), "97\u{2013}112"); |
| 1498 | assert_eq!(compress_pages("1--42"), "1\u{2013}42"); |
| 1499 | } |
| 1500 | |
| 1501 | #[test] |
| 1502 | fn reference_list_is_cited_only_and_sorted() { |
| 1503 | let mut b = bib(); |
| 1504 | // Cite three works by, or beginning with, Acemoglu plus Scott, out of order. |
| 1505 | let _ = b.cite(&["scott1976moral"]); |
| 1506 | let _ = b.cite(&["acemoglu2001colonial"]); |
| 1507 | let _ = b.cite(&["acemoglu2024simple"]); |
| 1508 | let list = b.reference_list(); |
| 1509 | assert_eq!(list.len(), 3); |
| 1510 | // Acemoglu solo (2024) sorts before Acemoglu-and-coauthors (2001); Scott last. |
| 1511 | assert_eq!(list[0].key, "acemoglu2024simple"); |
| 1512 | assert_eq!(list[1].key, "acemoglu2001colonial"); |
| 1513 | assert_eq!(list[2].key, "scott1976moral"); |
| 1514 | } |
| 1515 | |
| 1516 | #[test] |
| 1517 | fn empty_cited_list_is_empty() { |
| 1518 | let b = bib(); |
| 1519 | assert!(b.reference_list().is_empty()); |
| 1520 | } |
| 1521 | |
| 1522 | #[test] |
| 1523 | #[ignore] |
| 1524 | fn smoke_real_file() { |
| 1525 | let path = match std::env::var("REAL_BIB") { Ok(p) => p, Err(_) => return }; |
| 1526 | let src = std::fs::read_to_string(&path).expect("read real bib"); |
| 1527 | let b = Bibliography::parse(&src).expect("parse real bib"); |
| 1528 | println!("REAL entries parsed: {}", b.len()); |
| 1529 | let mut authored = 0usize; |
| 1530 | let mut yearless = 0usize; |
| 1531 | for e in &b.entries { |
| 1532 | if !e.credited().is_empty() { authored += 1; } |
| 1533 | if e.field("year").is_none() { yearless += 1; } |
| 1534 | } |
| 1535 | println!("with credited names: {} ; without year: {}", authored, yearless); |
| 1536 | // Spot-check a few keys against the oracle page 678. |
| 1537 | for key in ["scott1976moral", "kahneman1979prospect", "hirschman1970exit"] { |
| 1538 | if let Some(e) = b.entry(key) { |
| 1539 | println!("[{}] cite = {}", key, cite_label(e, "")); |
| 1540 | println!("[{}] ref = {}", key, format_reference(e, "").plain()); |
| 1541 | } |
| 1542 | } |
| 1543 | } |
| 1544 | |
| 1545 | #[test] |
| 1546 | fn author_count_tie_break() { |
| 1547 | // A solo work sorts before a coauthored one by the same first author. |
| 1548 | let src = r#" |
| 1549 | @book{a2012,author={Acemoglu, Daron and Robinson, James A.},title={Why Nations Fail},publisher={Crown},year={2012}} |
| 1550 | @book{a2023,author={Acemoglu, Daron and Johnson, Simon},title={Power and Progress},publisher={PublicAffairs},year={2023}} |
| 1551 | @techreport{a2024,author={Acemoglu, Daron},title={Simple},institution={NBER},year={2024}} |
| 1552 | "#; |
| 1553 | let b = match Bibliography::parse(src) { Ok(b) => b, Err(e) => panic!("{}", e) }; |
| 1554 | let all = b.all_references(); |
| 1555 | assert_eq!(all[0].key, "a2024"); // solo first |
| 1556 | assert_eq!(all[1].key, "a2023"); // Johnson before Robinson |
| 1557 | assert_eq!(all[2].key, "a2012"); |
| 1558 | } |
| 1559 | } |