From 04eca7c29ea914415fe62763eb538e73f71e102e Mon Sep 17 00:00:00 2001
From: Johannes de Wit <77152231+johannesdewit@users.noreply.github.com>
Date: Mon, 12 Feb 2024 00:10:46 +0100
Subject: [PATCH] Add ISO 639 part 3 validator for model field `lang`

---
 heidegger_index/migrations/0013_lemma_lang.py | 10 ++++++++--
 heidegger_index/models.py                     |  4 ++--
 heidegger_index/validators.py                 | 12 ++++++++++++
 requirements/production.txt                   |  3 ++-
 4 files changed, 24 insertions(+), 5 deletions(-)

diff --git a/heidegger_index/migrations/0013_lemma_lang.py b/heidegger_index/migrations/0013_lemma_lang.py
index fc05a73..71ecd41 100644
--- a/heidegger_index/migrations/0013_lemma_lang.py
+++ b/heidegger_index/migrations/0013_lemma_lang.py
@@ -1,9 +1,11 @@
-# Generated by Django 4.2.9 on 2024-02-09 08:50
+# Generated by Django 4.1 on 2024-02-11 23:08
 
 from django.db import migrations, models
+import heidegger_index.validators
 
 
 class Migration(migrations.Migration):
+
     dependencies = [
         ("heidegger_index", "0012_lemma_zeno"),
     ]
@@ -12,6 +14,10 @@ class Migration(migrations.Migration):
         migrations.AddField(
             model_name="lemma",
             name="lang",
-            field=models.CharField(max_length=3, null=True),
+            field=models.CharField(
+                max_length=3,
+                null=True,
+                validators=[heidegger_index.validators.validate_iso639_3_lang],
+            ),
         ),
     ]
diff --git a/heidegger_index/models.py b/heidegger_index/models.py
index 0f1960d..fe6023b 100644
--- a/heidegger_index/models.py
+++ b/heidegger_index/models.py
@@ -13,7 +13,7 @@
     slugify,
     contains_page_range,
 )
-from heidegger_index.validators import validate_gnd, validate_zeno
+from heidegger_index.validators import validate_gnd, validate_zeno, validate_iso639_3_lang
 
 
 class Work(models.Model):
@@ -80,7 +80,7 @@ class Lemma(models.Model):
     related = models.ManyToManyField("self", symmetrical=True)
     type = models.CharField(max_length=1, null=True, choices=LemmaType.list_choices())
     description = models.TextField(null=True)
-    lang = models.CharField(max_length=3, null=True)
+    lang = models.CharField(max_length=3, null=True, validators=[validate_iso639_3_lang])
 
     urn = models.URLField(
         MetadataType.URN.label,
diff --git a/heidegger_index/validators.py b/heidegger_index/validators.py
index f615760..d76224f 100644
--- a/heidegger_index/validators.py
+++ b/heidegger_index/validators.py
@@ -1,5 +1,7 @@
 from django.core.exceptions import ValidationError
 from .constants import MetadataType
+from iso639 import Lang
+from iso639.exceptions import InvalidLanguageValue
 
 import re
 
@@ -56,3 +58,13 @@ def validate_dk(value: int | str):
 
 def validate_zeno(value: str | int):
     validate_simple_regex(value, ZENO_REGEX, MetadataType.ZENO.label)
+
+
+def validate_iso639_3_lang(value: str):
+    try:
+        lang_code = Lang(value)
+    except InvalidLanguageValue as e:
+        raise ValidationError(e.msg)
+    
+    if lang_code.pt3 != value:
+        raise ValidationError(f'"{value}" is not a ISO 639 part 3 language code.')
\ No newline at end of file
diff --git a/requirements/production.txt b/requirements/production.txt
index 6544b71..a03b18e 100644
--- a/requirements/production.txt
+++ b/requirements/production.txt
@@ -20,4 +20,5 @@ aenum==3.*
 django-compressor==4.*
 django-libsass==0.*
 python-decouple==3.*
-lxml==5.*
\ No newline at end of file
+lxml==5.*
+iso639-lang
\ No newline at end of file