def migrate_glossaries(apps, schema_editor):
"""
Removes automatically created glossaries for source language.
These were wrongly created by 0127_fix_source_glossary since
0d8b564903518a313d4116ffe82d9c7bc31f7908 - it created blank repo.
"""
Component = apps.get_model("trans", "Component")
db_alias = schema_editor.connection.alias
for component in (Component.objects.using(db_alias).filter(
is_glossary=True,
repo="local:").prefetch_related("project", "source_language")):
repo_path = os.path.join(settings.DATA_DIR, "vcs",
component.project.slug, component.slug)
changed = False
for translation in component.translation_set.select_related(
"language"):
if translation.language_id == component.source_language_id:
continue
filename = os.path.join(repo_path, translation.filename)
if os.path.exists(filename):
continue
print(f"Adding missing {filename}")
TBXFormat.create_new_file(filename, translation.language.code, "")
store = TBXFormat(
filename,
language_code=translation.language.code,
source_language=component.source_language.code,
)
store.save()
changed = True
# Mark all strings a pending to be committed later
translation.unit_set.update(
pending=True,
details={"add_unit": True},
)
if changed:
repo = LocalRepository(repo_path)
with repo.lock:
repo.execute(["add", repo_path])
if repo.needs_commit():
repo.commit("Migrate glossary content")
def migrate_glossaries(apps, schema_editor): # noqa: C901
Project = apps.get_model("trans", "Project")
Language = apps.get_model("lang", "Language")
db_alias = schema_editor.connection.alias
projects = Project.objects.using(db_alias).all()
total = len(projects)
processed = 0
for processed, project in enumerate(projects):
component_slugs = set(
project.component_set.values_list("slug", flat=True))
percent = int(100 * processed / total)
print(
f"Migrating glossaries {percent}% [{processed}/{total}]...{project.name}"
)
glossaries = project.glossary_set.all()
try:
license = project.component_set.exclude(license="").values_list(
"license", flat=True)[0]
except IndexError:
license = ""
for glossary in glossaries:
if len(glossaries) == 1:
name = "Glossary"
slug = "glossary"
else:
name = f"Glossary: {glossary.name}"
slug = "glossary-{}".format(slugify(glossary.name))
base_name = name
base_slug = slug
# Create component
attempts = 0
while True:
if slug not in component_slugs:
component = create_glossary(project, name, slug, glossary,
license)
component_slugs.add(slug)
break
attempts += 1
name = f"{base_name} - {attempts}"
slug = f"{base_slug}-{attempts}"
repo_path = os.path.join(settings.DATA_DIR, "vcs", project.slug,
slug)
# Create VCS repository
repo = LocalRepository.from_files(repo_path, {})
# Migrate links
component.links.set(glossary.links.all())
# Create source translation
source_translation = component.translation_set.create(
language=glossary.source_language,
check_flags="read-only",
filename="",
plural=glossary.source_language.plural_set.filter(source=0)[0],
language_code=glossary.source_language.code,
)
source_units = {}
# Get list of languages
languages = Language.objects.filter(
term__glossary=glossary).distinct()
# Migrate ters
for language in languages:
base_filename = f"{language.code}.tbx"
filename = os.path.join(repo_path, base_filename)
is_source = language == source_translation.language
# Create translation object
if is_source:
translation = source_translation
else:
translation = component.translation_set.create(
language=language,
plural=language.plural_set.filter(source=0)[0],
filename=base_filename,
language_code=language.code,
)
# Create store file
TBXFormat.create_new_file(filename, language.code, "")
store = TBXFormat(filename, language_code=language.code)
id_hashes = set()
for position, term in enumerate(
glossary.term_set.filter(language=language)):
source = valid_chars_only(term.source)
target = valid_chars_only(term.target)
context = ""
# Store to the file
id_hash = calculate_hash(source, context)
offset = 0
while id_hash in id_hashes:
offset += 1
context = str(offset)
id_hash = calculate_hash(source, context)
id_hashes.add(id_hash)
if id_hash not in source_units:
source_units[
id_hash] = source_translation.unit_set.create(
context=context,
source=source,
target=source,
state=STATE_READONLY,
position=position,
num_words=len(source.split()),
id_hash=id_hash,
)
source_units[id_hash].source_unit = source_units[
id_hash]
source_units[id_hash].save()
store.new_unit(context, source, target)
# Migrate database
if is_source:
unit = source_units[id_hash]
unit.target = target
unit.save(update_fields=["target"])
else:
unit = translation.unit_set.create(
context=context,
source=source,
target=target,
state=STATE_TRANSLATED,
position=position,
num_words=len(source.split()),
id_hash=id_hash,
source_unit=source_units[id_hash],
)
# Adjust history entries (langauge and project should be already set)
term.change_set.update(
unit=unit,
translation=translation,
component=component,
)
store.save()
# Update translation hash
translation.revision = repo.get_object_hash(filename)
translation.save(update_fields=["revision"])
# Commit files
with repo.lock:
repo.execute(["add", repo_path])
if repo.needs_commit():
repo.commit("Migrate glossary content")
if total:
print(f"Migrating glossaries completed [{total}/{total}]")