Profile advisor
Paste a column of dates from a collection and the advisor says which conventions the data uses, before anything is parsed. Each string is still parsed on its own; the advisor only recommends the profile to parse it under, and writes that profile out for you to review and save. Nothing you paste is stored.
Or try a corpus sample: yearrange (museum catalogues) timetwister (US archives) Getty ULAN (artist dates) EDTF specification examples PMC (Paul Mellon Centre) outputs
yearrange (museum catalogues) 1,650 analysed · via the service
Recommendations
- 'early' (232), 'mid' (92) and 'late' (135) all occur: the default thirds are plausible; check a sample against the source's own definitions (quarters are also common: see profiles/catalogue-quarters.json).
- Numeric dates are day-first (1 strings with a first number over 12): keep dayMonthOrder dayFirst.
- 24 strings name a season: decide hemisphere and whether 'Winter 1776' is Dec-Feb or Jan-Mar (winter policy; profiles/archive-us.json).
- 319 strings name periods, dynasties or reigns: make sure the needed period tables are loaded (JsonNamedPeriodSource).
Suggested profile
Nothing in the data called for a change from the default profile.
{
"inherits": "default",
"name": "suggested",
"description": "Suggested by the profile advisor from 1650 strings: nothing differed from the default. Review before use."
}
Vocabulary
| Kind | Word or notation | Strings | For example |
|---|---|---|---|
| Band words | early | 226 | Early 20th century. c. 1884 (this item: First/Early edition) early 1790s |
| Band words | late | 99 | Edo period, Late, 1789-1868 Late 20th century. Late 20th C. |
| Band words | mid | 92 | Edo period, Middle, 1704-1789 Mid 20th century. Ca. mid 20th century. |
| Band words | later | 36 | [1765, printed later] [between 1790 and 1840, printed later] 1938 (later printing) |
| Band words | quarter | 13 | Third quarter 15th century. First quarter 13th century. Fourth quarter 13th century. |
| Band words | half | 12 | first half of the 16th century Second half of the 1850s. ca. first half 19th century |
| Band words | beginning/end of | 8 | Japanese, Meiji era, beginning of 20th century End of 16th century. Beginning of 16th century. |
| Band words | earlier | 6 | [between 1900 and 1940, from an earlier print] Earlier than 1932. Originally in 1811 or earlier, this is a re-carved edition probably in early 20th C. |
| Hedges | probably | 217 | probably 1773 Originally in Edo era. This re-carved edition was made probably in the early 20th century. Originally in Edo, this one is re-carved probably in 1930s. |
| Hedges | ca | 210 | Ca. 1870s. Ca. 1930. Ca. 1900-1910. |
| Hedges | c | 170 | c.1875 c.1890-1900 江戸時代_18c |
| Hedges | circa | 23 | 1821 (circa) 1628-1644 (circa) circa 1876-1881 |
| Hedges | around | 16 | Originally in 1858, this was a re-carved edition published in around 1900. Originally carved in Edo era. This re-carved edition was made in around 1950s. This edition was printed probably around 1890s. |
| Hedges | about | 6 | 1823-1829 (about) 1840 (about) 1821 (ninth month, about) |
| Hedges | v | 1 | [1862 (seal date: Boar V)] |
| Century notations | word | 269 | Late 20th century. 20th century. Early 20th century. |
| Century notations | 19th c. | 103 | 19thC Late 20th C. Originally in Edo era, this one is re-carved edition in late 20th C. |
| Century notations | roman | 11 | Originally in 1855, this re-carved edition was made before W.W.II. Originally published in 1833/1834. This edition published probably before world war II in 20th century. 1930 ... Printing before W.W.II. |
| Century notations | 1800s | 10 | Ca. mid 1700s. Originally published in Edo era. Re-carved edition made on ca. 1900s. Originally in Edo era, this is a re-carved edition made in ca. 1900s. |
| Other languages | fr | 1 | [1862 (seal date: Boar V)] |
Shapes
| Shape | Strings | For example | Why it matters |
|---|---|---|---|
| Numeric dates | 3 | 1904 (1/7/1904) 1904 (23/3/1904) 02.04.1979. | Day, month and year as numbers. |
| Day-first evidence | 1 | 1904 (23/3/1904) | A first number over 12, so it can only be a day. |
| Month-first evidence | 0 | A second number over 12, so the first is the month. | |
| Two-digit years | 0 | Need a window to become four-digit years. | |
| Bracketed | 84 | [between 1870 and 1890] [1904 or 1905] [187-] | Inferred dates in MARC style. |
| Seasons | 24 | 1847 (spring) spring 1864 summer 1922 | Need a hemisphere and a winter policy. |
| Undated | 88 | Originally in Edo era, this re-carved edition was made in unknown date. 1859, 2nd month Originally in Edo era, this one is a re-carved edition made in unknown date. | The text says there is no date. |
| Open ends | 97 | [187-] Originally in Edo era, this re-carved edition was made in unknown date. [18--] | Before, after, present, a trailing dash. |
| Decades | 207 | Ca. 1870s. 1870s. Ca. 1900-1910s. | 1950s and the like; band words apply to them too. |
| Period words | 319 | Japanese, Edo period Edo period, Late, 1789-1868 Edo period, 1615-1868 | Dynasties, reigns and named periods need period tables. |
The same advice from code
The advisor is a library call and an endpoint on the dates API; the response carries these counts, the examples, the recommendations, the suggested profile and a Markdown rendering.
POST https://dates.linkedart.net/v1/advise
Content-Type: text/plain
early 19th century
ca. 1850
7/4/1776
var report = ProfileAdvisor.Analyse(File.ReadLines("dates.txt"));
Console.WriteLine(report.ToMarkdown());
File.WriteAllText("profiles/mine.json", report.SuggestedProfileJson("mine").ToJsonString(ParseProfile.JsonOptions));
Or from the repository: dotnet run --project tools/Corpus -- advise dates.txt.