gedcomkit 0.1.18__tar.gz → 0.1.20__tar.gz

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/CHANGELOG.md

@@ -7,6 +7,103 @@ rules.
 
 ## [Unreleased]
 
+## [0.1.20] - 2026-11-09
+
+### Fixed
+
+- The date reader reads a calendar named after a qualifier or a range
+  word, as both grammars write it: `INT @#DJULIAN@ 2 MAR 1700 (…)`,
+  `BEF @#DJULIAN@ 2 JAN 1700`, `ABT JULIAN 1700`, `BET @#DJULIAN@ 1700 AND
+  @#DJULIAN@ 1710`. Such a date was left unparsed before, though
+  `to_gedcom_5_5_1` writes exactly this form, so what it wrote did not read
+  back. A range whose ends name different calendars, or whose unnamed end
+  (Gregorian) disagrees with a named one, stays unparsed.
+
+## [0.1.19] - 2026-10-17
+
+### Added
+
+- `view::EventView::date_phrase` (`datePhrase` in the serialized view, so
+  in both bindings and the TypeScript definitions): the `PHRASE` beneath an
+  event's `DATE`, exactly as written. A GEDCOM 7 date with an empty payload
+  and a `PHRASE`, or a date with a phrase beside it, showed nothing of the
+  phrase before. A 5.5.1 phrase is written inside the payload, which `date`
+  already holds. The Python stub gains an `EventView` typed dict naming
+  every key of an event.
+- The date reader reads ISO 8601's year and month: `1834-02` is February
+  1834, as `1834-02-10` was already 10 February 1834. Only four digits, a
+  hyphen and two digits are read, and only when the month is no more than
+  the year's last two digits: `1808-10`, `1711-12` and `2010-11` are as
+  often a span of years or a dual year as a month, and stay unparsed. So
+  does every month of a year ending in 00 (`1800-01` to `1800-12`), and
+  the later months of a year ending in 01 to 11 (`2005-06` to `2005-12`).
+
+### Changed
+
+- `validate::inspect` no longer reports two events out of order when an
+  approximate date's leeway could put them in order, which is how
+  `conflicting-vitals` already read them. A birth `ABT 1854` with a
+  christening in 1853 was a `christened-before-born` question; it is now
+  no finding. The same holds for `died-before-born`, `buried-before-died`,
+  `event-after-death`, `event-before-birth`, `married-before-born`,
+  `married-after-died` and `child-born-after-mother-died`, each of which
+  now reports only what is out of order beyond the leeway, as a
+  contradiction (a question for an `EVEN` or `FACT`, as before). Within the
+  leeway no other check fires in its place: a marriage in 1849 for somebody
+  born about 1850 is not `married-as-a-child`. Bounds (`BEF`, `AFT`, `BET`)
+  are read as the ranges they admit, unchanged.
+- `convert::to_version_5_5_1` turns a name type back into 5.5.1's
+  spelling: `OTHER` with a one-line `PHRASE` becomes the phrase's words,
+  and an enumerated value is written in lower case (`BIRTH` becomes
+  `birth`). `OTHER` whose words are themselves an enumerated value (the
+  phrase `birth`) stays `other`, its `PHRASE` kept and reported, since the
+  words alone would convert back up as `BIRTH`. A `PHRASE` beside any other
+  value is kept and reported.
+- `DatePoint::from_ymd` refuses a day its month never has (30 February,
+  31 April, day 0). The point is written with no calendar escape, which
+  makes it Gregorian, so 29 February is taken only in a Gregorian leap
+  year: 2000, not 1900.
+
+### Fixed
+
+- A range with an end the date reader cannot read is no longer read in
+  part. `FROM ABT 1924 TO 1959` was read as `TO 1959`, dropping the start
+  without a word, and `BET 1950 AND JULIAN DEC 302` as a range with only
+  its first end; `FROM Infant` was a period with neither end. Each is now
+  `DateKind::Unparsed`, keeps its text as the phrase, and still offers its
+  year to sort by; `to_version_7` keeps it whole as an empty `DATE` with a
+  `PHRASE`. A range whose ends are both dates reads as before.
+- The date reader no longer accepts a day its month does not have.
+  `1834-02-30` and `30 FEB 1834` were read as 30 February 1834; they are
+  now unparsed, as is `29 FEB 1834`. Leap years follow the payload's
+  calendar: `@#DJULIAN@ 29 FEB 1900` is read and `@#DGREGORIAN@ 29 FEB
+  1900` is not. A payload with no calendar escape keeps the Julian leap
+  day, since files often leave the escape off a Julian date, so
+  `29 FEB 1900` is still read; ISO 8601's `1900-02-29` is not, since ISO
+  8601 is Gregorian by definition. A dual year's leap day is its second
+  year's, so `29 FEB 1711/12` and `@#DJULIAN@ 29 FEB 1711/12` are read. A
+  Roman or unrecognized calendar is held to the Julian month lengths, and
+  no Hebrew or French Republican month has more than 30 days. A numeric
+  date with a month or day out of range (`1834-13-05`, `13/05/1834`) is
+  unparsed too; before, the piece out of range was dropped and the rest
+  read (`1834-13-05` as the year 1834).
+- `convert::to_version_5_5_1` no longer writes `INT  (…)`, an interpreted
+  date with no date, for a version 7 `DATE` that is empty beside its
+  `PHRASE`; 5.5.1 does not allow it. The phrase is written as 5.5.1's date
+  phrase, `(FROM ABT 1924 TO 1959)`, or as the date itself when it is
+  already one in 5.5.1's grammar that version 7 cannot spell, so a dual
+  year (`1740/41`) comes back as the date it was.
+- `convert::to_version_7` / `to_version_7_with` convert a personal name's
+  type (`INDI`.`NAME`.`TYPE`), which 5.5.1 writes as free text and version
+  7 as an enumeration. A value that is one of version 7's words in any case
+  (`aka`, `birth`, `immigrant`, `maiden`, `married`, `professional`,
+  `other`) becomes that word in upper case; any other becomes `OTHER`, with
+  the text kept verbatim as a `PHRASE`, the way a role has been converted
+  since 0.1.13. Before, `TYPE aka` was carried into the 7.0 document as it
+  stood, where it is not valid. A type that already has a `PHRASE`, is
+  empty, or has a continuation line with substructures is kept and
+  reported.
+
 ## [0.1.18] - 2026-10-16
 
 ### Added

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/Cargo.lock

@@ -35,7 +35,7 @@ dependencies = [
 
 [[package]]
 name = "gedcomkit"
-version = "0.1.18"
+version = "0.1.20"
 dependencies = [
  "flate2",
  "serde",

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/Cargo.toml

@@ -3,7 +3,7 @@
 # was checked on crates.io, npm, and PyPI before the initial release.
 [package]
 name = "gedcomkit"
-version = "0.1.18"
+version = "0.1.20"
 description = "A byte-preserving GEDCOM document model: decoding, parsing, readings, version conversion, plausibility checks, and the GEDZIP container, for GEDCOM 5.5 through 7.x."
 edition = "2024"
 # The true minimum, not the workspace's toolchain: a parser library is a leaf

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gedcomkit
-Version: 0.1.18
+Version: 0.1.20
 Classifier: Development Status :: 3 - Alpha
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/README.md

@@ -42,12 +42,8 @@ npm install gedcomkit
 2. Run the gates and release dry run:
 
    ```bash
-   cargo fmt --all --check
-   cargo clippy --all-features --all-targets -- -D warnings
-   cargo test --all-features
-   cargo test --no-default-features
-   cargo check --manifest-path bindings/gedcomkit-wasm/Cargo.toml --target wasm32-unknown-unknown
-   cargo check --manifest-path bindings/gedcomkit-py/Cargo.toml
+   cargo ci
+   cargo +1.88 ci test
    cargo publish --locked --dry-run
    bash scripts/check-release-tag.sh vX.Y.Z
    ```

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/bindings/gedcomkit-py/Cargo.lock

@@ -35,7 +35,7 @@ dependencies = [
 
 [[package]]
 name = "gedcomkit"
-version = "0.1.18"
+version = "0.1.20"
 dependencies = [
  "flate2",
  "serde",
@@ -43,7 +43,7 @@ dependencies = [
 
 [[package]]
 name = "gedcomkit-py"
-version = "0.1.18"
+version = "0.1.20"
 dependencies = [
  "gedcomkit",
  "pyo3",

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/bindings/gedcomkit-py/Cargo.toml

@@ -2,7 +2,7 @@
 # abi3 keeps it to one wheel per platform across Python versions.
 [package]
 name = "gedcomkit-py"
-version = "0.1.18"
+version = "0.1.20"
 description = "Python bindings for gedcomkit: an archival-grade GEDCOM library — byte-preserving round-trip, encoding detection, GEDCOM 5.5 through 7.x, GEDZIP."
 edition = "2024"
 rust-version = "1.88"

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/bindings/gedcomkit-py/gedcomkit.pyi

@@ -19,6 +19,29 @@ class RecordSpec(TypedDict):
     value: NotRequired[str | None]
     children: NotRequired[list[RecordSpec]]
 
+class EventView(TypedDict):
+    """One entry of an individual's or a family's ``events``.
+
+    ``date`` is the DATE payload exactly as written. ``datePhrase`` is the
+    GEDCOM 7 PHRASE beneath that DATE, or ``None``: beside a date, or the
+    only text an empty 7.0 DATE has. A 5.5.1 phrase is part of ``date``.
+    """
+
+    tag: str
+    label: str
+    date: str | None
+    datePhrase: str | None
+    dateDisplay: str | None
+    year: int | None
+    sortKey: int | None
+    place: str | None
+    age: str | None
+    cause: str | None
+    attribute: bool
+    value: str | None
+    citations: list[str]
+    path: list[int]
+
 class WrappedLine(TypedDict):
     record: str
     path: list[int]
@@ -42,8 +65,10 @@ class Gedcom:
     def content_records(self) -> int: ...
     def records(self) -> list[dict[str, Any]]: ...
     def record(self, xref: str) -> dict[str, Any]: ...
-    def individuals(self) -> list[dict[str, Any]]: ...
-    def families(self) -> list[dict[str, Any]]: ...
+    def individuals(self) -> list[dict[str, Any]]:
+        """Every individual's view; each of its ``events`` is an ``EventView``."""
+    def families(self) -> list[dict[str, Any]]:
+        """Every family's view; each of its ``events`` is an ``EventView``."""
     def validate(self) -> list[dict[str, Any]]:
         """Plausibility findings, each with ``record``, ``check``, ``severity``
         and ``message``. ``check`` is the core's stable rule name, such as

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/bindings/gedcomkit-py/src/lib.rs

@@ -224,6 +224,8 @@ impl Gedcom {
     }
 
     /// Every individual, read for display: names, events, families, IDs.
+    /// Each event has its `date` payload as written and, beside it,
+    /// `datePhrase`: the GEDCOM 7 `PHRASE` under the date, or `None`.
     fn individuals(&self, py: Python<'_>) -> PyResult<Vec<Py<PyAny>>> {
         self.document
             .records_of("INDI")
@@ -231,7 +233,7 @@ impl Gedcom {
             .collect()
     }
 
-    /// Every family record's view.
+    /// Every family record's view, its events shaped as `individuals`' are.
     fn families(&self, py: Python<'_>) -> PyResult<Vec<Py<PyAny>>> {
         self.document
             .records_of("FAM")

{gedcomkit-0.1.18 → gedcomkit-0.1.20}/src/build.rs

@@ -316,7 +316,7 @@ impl Node {
     /// beside a space loses that space on re-joining. Failing that — a run
     /// of spaces wider than the limit, say — the cut falls back to the
     /// latest legal byte boundary instead, which is documented behavior of
-    /// [`cut`] this call reuses, not a special case here.
+    /// the private `cut` this call reuses, not a special case here.
     ///
     /// The rendered line is measured in UTF-8 bytes, including level,
     /// optional cross-reference, tag, delimiters, and a terminator reserved