From 3708d06b34f1be0f7ce9d39ff8993b7760f77bc2 Mon Sep 17 00:00:00 2001
From: Luther Tychonievich <tychonievich@gmail.com>
Date: Mon, 10 Feb 2025 16:00:58 -0600
Subject: [PATCH] Update how tables with URIs in them are parsed (#590)

* Update how tables with URIs in them are parsed

Resolves #574

* Add the extracted files this PR changes

* Fix over-eager change of tag names to enum-TAG
---
 build/uri-def.py                      | 11 ++++++++---
 extracted-files/enumerationsets.tsv   |  4 ++--
 extracted-files/tags/INDI-RELI        |  6 ++++++
 extracted-files/tags/INDI-TITL        |  6 ++++++
 extracted-files/tags/RELI             |  6 ------
 extracted-files/tags/TITL             |  6 ------
 extracted-files/tags/enumset-EVENATTR |  4 ++--
 7 files changed, 24 insertions(+), 19 deletions(-)

diff --git a/build/uri-def.py b/build/uri-def.py
index 14b9fad5..ffc1a83d 100644
--- a/build/uri-def.py
+++ b/build/uri-def.py
@@ -251,7 +251,9 @@ def find_descriptions(txt, g7, ssp):
         if header.startswith('Fam'): pfx = 'FAM-'
         if header.startswith('Indi'): pfx = 'INDI-'
         for tag, name, desc in re.findall(r'`([A-Z_0-9]+)` *\| *([^|\n]*?) *\| *([^|\n]*[^ |\n]) *', table.group(2)):
-            if '<br' in name: name = name[:name.find('<br')]
+            if '<br' in name:
+                tag = name[name.find('`g7:')+4:name.rfind('`')]
+                name = name[:name.find('<br')]
             if tag not in g7: tag = pfx+tag
             if tag not in g7:
                 raise Exception('Found table for '+tag+' but no section or structure')
@@ -270,6 +272,7 @@ def find_enum_by_link(txt, enums, tagsets):
             # 'g7:INDI-FACT',
             # 'g7:FAM-FACT',
         # ))         ## do not do for enumset-EVEN 
+    enum_prefix = {k[k.find('enum-')+5:] for e in enums.values() for k in e }
     for sect in re.finditer(r'# *`(g7:enumset-[^`]*)`[\s\S]*?\n#', txt):
         if '[Events]' in sect.group(0):
             key = sect.group(1).replace('`','').replace('.','-')
@@ -277,7 +280,8 @@ def find_enum_by_link(txt, enums, tagsets):
                 if 'Event' in k:
                     enums.setdefault(key, [])
                     for tag in tagsets[k]:
-                        tag = tag.replace('INDI-','enum-').replace('FAM-','enum-')
+                        if tag.startswith('INDI-') and tag[5:] in enum_prefix: tag = 'enum-'+tag[5:]
+                        if tag.startswith('FAM-') and tag[4:] in enum_prefix: tag = 'enum-'+tag[4:]
                         tag = 'g7:'+tag
                         if tag in enums[key]: continue
                         enums[key].append(tag)
@@ -287,7 +291,8 @@ def find_enum_by_link(txt, enums, tagsets):
                 if 'Attribute' in k:
                     enums.setdefault(key, [])
                     for tag in tagsets[k]:
-                        tag = tag.replace('INDI-','enum-').replace('FAM-','enum-')
+                        if tag.startswith('INDI-') and tag[5:] in enum_prefix: tag = 'enum-'+tag[5:]
+                        if tag.startswith('FAM-') and tag[4:] in enum_prefix: tag = 'enum-'+tag[4:]
                         tag = 'g7:'+tag
                         if tag in enums[key]: continue
                         enums[key].append(tag)
diff --git a/extracted-files/enumerationsets.tsv b/extracted-files/enumerationsets.tsv
index bd103f15..3eb80a8f 100644
--- a/extracted-files/enumerationsets.tsv
+++ b/extracted-files/enumerationsets.tsv
@@ -75,9 +75,9 @@ https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/NATI
 https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/NMR
 https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/OCCU
 https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/PROP
-https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/RELI
+https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/INDI-RELI
 https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/SSN
-https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/TITL
+https://gedcom.io/terms/v7/enumset-EVENATTR	https://gedcom.io/terms/v7/INDI-TITL
 https://gedcom.io/terms/v7/enumset-MEDI	https://gedcom.io/terms/v7/enum-AUDIO
 https://gedcom.io/terms/v7/enumset-MEDI	https://gedcom.io/terms/v7/enum-BOOK
 https://gedcom.io/terms/v7/enumset-MEDI	https://gedcom.io/terms/v7/enum-CARD
diff --git a/extracted-files/tags/INDI-RELI b/extracted-files/tags/INDI-RELI
index 1d4b82d3..dcbefbcb 100644
--- a/extracted-files/tags/INDI-RELI
+++ b/extracted-files/tags/INDI-RELI
@@ -11,6 +11,9 @@ standard tag: 'RELI'
 specification:
   - Religion
   - An [Individual Attribute]. See also `INDIVIDUAL_ATTRIBUTE_STRUCTURE`.
+  - religion
+  - A religious denomination to which a person is affiliated or for which a record
+    applies.
 
 label: 'Religion'
 
@@ -41,5 +44,8 @@ substructures:
 superstructures:
   "https://gedcom.io/terms/v7/record-INDI": "{0:M}"
 
+value of:
+  - "https://gedcom.io/terms/v7/enumset-EVENATTR"
+
 contact: "https://gedcom.io/community/"
 ...
diff --git a/extracted-files/tags/INDI-TITL b/extracted-files/tags/INDI-TITL
index b54ba5cf..d16a3e99 100644
--- a/extracted-files/tags/INDI-TITL
+++ b/extracted-files/tags/INDI-TITL
@@ -11,6 +11,9 @@ standard tag: 'TITL'
 specification:
   - Title
   - An [Individual Attribute]. See also `INDIVIDUAL_ATTRIBUTE_STRUCTURE`.
+  - title
+  - A formal designation used by an individual in connection with positions of
+    royalty or other social status, such as Grand Duke.
 
 label: 'Title'
 
@@ -41,5 +44,8 @@ substructures:
 superstructures:
   "https://gedcom.io/terms/v7/record-INDI": "{0:M}"
 
+value of:
+  - "https://gedcom.io/terms/v7/enumset-EVENATTR"
+
 contact: "https://gedcom.io/community/"
 ...
diff --git a/extracted-files/tags/RELI b/extracted-files/tags/RELI
index 9650917d..9e22c0b0 100644
--- a/extracted-files/tags/RELI
+++ b/extracted-files/tags/RELI
@@ -12,9 +12,6 @@ specification:
   - Religion
   - A religious denomination associated with the event or attribute described by
     the superstructure.
-  - religion
-  - A religious denomination to which a person is affiliated or for which a record
-    applies.
 
 label: 'Religion'
 
@@ -75,8 +72,5 @@ superstructures:
   "https://gedcom.io/terms/v7/SSN": "{0:1}"
   "https://gedcom.io/terms/v7/WILL": "{0:1}"
 
-value of:
-  - "https://gedcom.io/terms/v7/enumset-EVENATTR"
-
 contact: "https://gedcom.io/community/"
 ...
diff --git a/extracted-files/tags/TITL b/extracted-files/tags/TITL
index e32a16cb..b68e0e6e 100644
--- a/extracted-files/tags/TITL
+++ b/extracted-files/tags/TITL
@@ -37,9 +37,6 @@ specification:
     the `SOURCE_RECORD` substructures `AUTH`, `PUBL`, `REPO`, and so on. In such
     cases, the entire citation text may be presented as the payload of the
     `SOUR`.`TITL`.
-  - title
-  - A formal designation used by an individual in connection with positions of
-    royalty or other social status, such as Grand Duke.
 
 label: 'Title'
 
@@ -52,8 +49,5 @@ superstructures:
   "https://gedcom.io/terms/v7/OBJE": "{0:1}"
   "https://gedcom.io/terms/v7/record-SOUR": "{0:1}"
 
-value of:
-  - "https://gedcom.io/terms/v7/enumset-EVENATTR"
-
 contact: "https://gedcom.io/community/"
 ...
diff --git a/extracted-files/tags/enumset-EVENATTR b/extracted-files/tags/enumset-EVENATTR
index e38d6ce7..f93d79f4 100644
--- a/extracted-files/tags/enumset-EVENATTR
+++ b/extracted-files/tags/enumset-EVENATTR
@@ -50,9 +50,9 @@ enumeration values:
   - "https://gedcom.io/terms/v7/NMR"
   - "https://gedcom.io/terms/v7/OCCU"
   - "https://gedcom.io/terms/v7/PROP"
-  - "https://gedcom.io/terms/v7/RELI"
+  - "https://gedcom.io/terms/v7/INDI-RELI"
   - "https://gedcom.io/terms/v7/SSN"
-  - "https://gedcom.io/terms/v7/TITL"
+  - "https://gedcom.io/terms/v7/INDI-TITL"
 
 contact: "https://gedcom.io/community/"
 ...