Add authority and availability fields

Xennis · Jun 6, 2024 · 0872ea7 · 0872ea7
1 parent a3683d4
commit 0872ea7
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 2 deletions.
diff --git a/README.md b/README.md
@@ -63,6 +63,8 @@ Call the attributes, for example
 | edition_foreign_languages | `//body/div[@type='edition']//foreign/@xml:lang`                               |
 | edition_language          | `//body/div[@type='edition']/@xml:lang`                                        |
 | idno                      | `//teiHeader/fileDesc/publicationStmt/idno`                                    |
+| authority                 | `//teiHeader/fileDesc/publicationStmt/authority`                               |
+| availability              | `//teiHeader/fileDesc/publicationStmt/availability`                            |
 | languages                 | `//teiHeader/profileDesc/langUsage/language`                                   |
 | material                  | `//teiHeader/fileDesc/sourceDesc/msDesc/physDesc/objectDesc//support/material` |
 | origin_dates              | `//teiHeader/fileDesc/sourceDesc/msDesc/history/origin/origDate`               |

diff --git a/epidoc/api.py b/epidoc/api.py
@@ -1,3 +1,5 @@
+from typing import Optional
+
 from bs4 import BeautifulSoup
 
 from .body import _Edition, _Head
@@ -9,6 +11,8 @@ class EpiDoc:
 
     title = None
     idno: dict[str, str] = {}
+    authority: Optional[str] = None
+    availability: Optional[str] = None
     material = None
     origin_dates: list[str] = []
     origin_place: dict[str, str] = {}
@@ -26,6 +30,8 @@ def create(
         cls,
         title,
         idno,
+        authority=None,
+        availability=None,
         material=None,
         origin_dates=None,
         origin_place=None,
@@ -42,6 +48,10 @@ def create(
         h.title = title
         h.idno = idno
         h.material = material
+        if authority is not None:
+            h.authority = authority
+        if availability is not None:
+            h.availability = availability
         if origin_dates is not None:
             h.origin_dates = origin_dates
         if origin_place is not None:
@@ -78,13 +88,26 @@ def loads(s):
     filedesc = teiheader.filedesc
     doc.title = filedesc.titlestmt.title.getText()
     idnos = {}
-    for idno in filedesc.publicationstmt.find_all("idno"):
+    publication_stmt = filedesc.publicationstmt
+    for idno in publication_stmt.find_all("idno"):
         typ = _normalize(idno.attrs.get("type"))
         value = _normalize(idno.getText())
         if not value:
             continue
         idnos[typ] = value
     doc.idno = idnos
+    authority = publication_stmt.find("authority")
+    if authority:
+        doc.authority = _normalized_get_text(authority)
+    availability = publication_stmt.find("availability")
+    if availability:
+        availability_text = _normalized_get_text(availability)
+        license = availability.find("ref", type="license")
+        if license:
+            license_target = license.attrs.get("target")
+            if license_target:
+                availability_text += f" {license_target}"
+        doc.availability = availability_text
 
     msdesc = filedesc.sourcedesc.msdesc
     if msdesc:

diff --git a/epidoc/normalize.py b/epidoc/normalize.py
@@ -14,7 +14,7 @@ def _normalize(v: T) -> T:
 def _normalized_get_text(raw):
     if not raw:
         return None
-    parsed = raw.getText().strip()
+    parsed = raw.getText().strip().replace("\n", "")
     return parsed if parsed else None
 
 

diff --git a/tests/test_api.py b/tests/test_api.py
@@ -15,6 +15,8 @@
 def assert_epi_doc(test, want, actual):
     test.assertEqual(want.title, actual.title, msg="title")
     test.assertEqual(want.idno, actual.idno, msg="idno")
+    test.assertEqual(want.authority, actual.authority, msg="authority")
+    test.assertEqual(want.availability, actual.availability, msg="availability")
     test.assertEqual(want.material, actual.material, msg="material")
     test.assertEqual(want.origin_dates, actual.origin_dates, msg="origin_dates")
     test.assertEqual(want.origin_place, actual.origin_place, msg="origin_place")
@@ -24,6 +26,8 @@ def assert_epi_doc(test, want, actual):
     test.assertEqual(want.commentary, actual.commentary, msg="commentary")
     test.assertEqual(want.edition_language, actual.edition_language, msg="edition_language")
     test.assertEqual(want.edition_foreign_languages, actual.edition_foreign_languages, msg="edition_foreign_languages")
+    test.assertEqual(want.reprint_from, actual.reprint_from, msg="reprint_from")
+    test.assertEqual(want.reprint_in, actual.reprint_in, msg="reprint_in")
 
 
 class TestLoad(unittest.TestCase):
@@ -34,13 +38,16 @@ def test_all(self):
                 EpiDoc.create(
                     title="P.Heid.Arab. I p. 6-7",
                     idno={"filename": "pap(23new)", "tm": "106496"},
+                    authority="The Arabic Papyrology Database",
+                    availability="© The Arabic Papyrology Database. This work is licensed under a Creative Commons Attribution 3.0 License. http://creativecommons.org/licenses/by/3.0/",
                 ),
             ),
             (
                 os.path.join("apis", "yale.apis.0000540000.xml"),
                 EpiDoc.create(
                     title="Receipt, Roman CE ii (ca. 162) [BPG]",
                     idno={"apisid": "yale.apis.0000540000", "controlno": "(cty)54", "hgv": "20671", "tm": "20671"},
+                    authority="APIS",
                     origin_dates=[{"notafter": "0199", "notbefore": "0100", "text": "Roman CE ii (ca. 162) [BPG]"}],
                     terms=[{"text": "Receipt"}, {"text": "Papyri"}],
                     languages={"en": "English", "grc": "Greek"},
@@ -57,8 +64,11 @@ def test_all(self):
                         "hgv": "114844",
                         "tm": "114844",
                     },
+                    authority="Duke Collaboratory for Classics Computing (DC3)",
+                    availability="© Duke Databank of Documentary Papyri. This work is licensed under a Creative Commons Attribution 3.0 License. http://creativecommons.org/licenses/by/3.0/",
                     languages={"en": "English", "la": "Latin"},
                     edition_language="la",
+                    reprint_in=["p.ital;2;38/41"],
                 ),
             ),
             (
@@ -71,6 +81,8 @@ def test_all(self):
                         "hgv": "697551",
                         "tm": "697551",
                     },
+                    authority="Duke Collaboratory for Classics Computing (DC3)",
+                    availability="© Duke Databank of Documentary Papyri. This work is licensed under a          Creative          Commons Attribution 3.0 License. http://creativecommons.org/licenses/by/3.0/",
                     languages={"en": "English", "grc": "Greek"},
                     edition_language="grc",
                 ),
@@ -86,6 +98,8 @@ def test_all(self):
                         "ldab": "5148",
                         "tm": "26761",
                     },
+                    authority="Digital Corpus of Literary Papyri",
+                    availability="© Digital Corpus of Literary Papyri. This work is licensed under a Creative Commons Attribution 3.0 License. http://creativecommons.org/licenses/by/3.0/",
                     material="papyrus",
                     origin_dates=[
                         {
@@ -132,6 +146,8 @@ def test_all(self):
                         "ldab": "135858",
                         "tm": "135858",
                     },
+                    authority="Digital Corpus of Literary Papyri",
+                    availability="© Digital Corpus of Literary Papyri. This work is licensed under a Creative Commons Attribution 3.0 License. http://creativecommons.org/licenses/by/3.0/",
                     material="parchment",
                     origin_dates=[
                         {