def rewrite_file_paths(path, test_files_dir):
"""Load the import CSV, prepend pytest tmpdir to basename_orig
"""
headers, rowds, csv_errs = csvfile.make_rowds(fileio.read_csv(path))
for rowd in rowds:
src = os.path.join(test_files_dir, rowd['basename_orig'])
rowd['basename_orig'] = src
headers, rows = csvfile.make_rows(rowds)
fileio.write_csv(path, headers, rows)
def test_read_csv():
# prep
if os.path.exists(CSV_PATH):
os.remove(CSV_PATH)
with open(CSV_PATH, 'w') as f:
f.write(CSV_FILE)
# test
expected = CSV_ROWS
expected.insert(0, CSV_HEADERS)
assert fileio.read_csv(CSV_PATH) == expected
# cleanup
if os.path.exists(CSV_PATH):
os.remove(CSV_PATH)
def test_read_csv(tmpdir):
CSV_PATH = str(tmpdir / 'write_csv.csv')
# prep
if os.path.exists(CSV_PATH):
os.remove(CSV_PATH)
with open(CSV_PATH, 'w') as f:
f.write(CSV_FILE)
# test
expected = CSV_ROWS
expected.insert(0, CSV_HEADERS)
assert fileio.read_csv(CSV_PATH) == expected
# cleanup
if os.path.exists(CSV_PATH):
os.remove(CSV_PATH)
def check_csv(csv_path, cidentifier, vocabs_path):
"""Load CSV, validate headers and rows
Results dict includes:
- 'passed'
- 'headers'
- 'rowds'
- 'header_errs'
- 'rowds_errs'
@param csv_path: Absolute path to CSV data file.
@param cidentifier: Identifier
@param vocabs_path: Absolute path to vocab dir
@param session: requests.session object
@returns: nothing
"""
logging.info('Checking CSV file')
passed = False
headers,rowds = csvfile.make_rowds(fileio.read_csv(csv_path))
for rowd in rowds:
rowd['identifier'] = identifier.Identifier(rowd['id'])
logging.info('%s rows' % len(rowds))
model = Checker._guess_model(rowds)
module = Checker._get_module(model)
vocabs = Checker._get_vocabs(module)
header_errs,rowds_errs = Checker._validate_csv_file(
module, vocabs, headers, rowds
)
if (not header_errs) and (not rowds_errs):
passed = True
logging.info('ok')
else:
logging.error('FAIL')
return {
'passed': passed,
'headers': headers,
'rowds': rowds,
'header_errs': header_errs,
'rowds_errs': rowds_errs,
}
def register_entity_ids(csv_path, cidentifier, idservice_client, dryrun=True):
"""
@param csv_path: Absolute path to CSV data file.
@param cidentifier: Identifier
@param idservice_client: idservice.IDServiceCrequests.session object
@param register: boolean Whether or not to register IDs
@returns: nothing
"""
logging.info('-----------------------------------------------')
logging.info('Reading %s' % csv_path)
headers,rowds = csvfile.make_rowds(fileio.read_csv(csv_path))
logging.info('%s rows' % len(rowds))
logging.info('Looking up already registered IDs')
csv_eids = [rowd['id'] for rowd in rowds]
status1,reason1,registered,unregistered = idservice_client.check_eids(cidentifier, csv_eids)
logging.info('%s %s' % (status1,reason1))
if status1 != 200:
raise Exception('%s %s' % (status1,reason1))
num_unregistered = len(unregistered)
logging.info('%s IDs to register.' % num_unregistered)
if unregistered and dryrun:
logging.info('These IDs would be registered if not --dryrun')
for n,eid in enumerate(unregistered):
logging.info('| %s/%s %s' % (n, num_unregistered, eid))
elif unregistered:
logging.info('Registering IDs')
for n,eid in enumerate(unregistered):
logging.info('| %s/%s %s' % (n, num_unregistered, eid))
status2,reason2,created = idservice_client.register_eids(cidentifier, unregistered)
logging.info('%s %s' % (status2,reason2))
if status2 != 201:
raise Exception('%s %s' % (status2,reason2))
logging.info('%s registered' % len(created))
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
def import_files(csv_path, cidentifier, vocabs_path, git_name, git_mail, agent, log_path=None, dryrun=False):
"""Adds or updates files from a CSV file
TODO how to handle excluded fields like XMP???
@param csv_path: Absolute path to CSV data file.
@param cidentifier: Identifier
@param vocabs_path: Absolute path to vocab dir
@param git_name: str
@param git_mail: str
@param agent: str
@param log_path: str Absolute path to addfile log for all files
@param dryrun: boolean
"""
logging.info('batch import files ----------------------------')
# TODO hard-coded model name...
model = 'file'
csv_dir = os.path.dirname(csv_path)
logging.debug('csv_dir %s' % csv_dir)
# TODO this still knows too much about entities and files...
entity_class = identifier.class_for_name(
identifier.MODEL_CLASSES['entity']['module'],
identifier.MODEL_CLASSES['entity']['class']
)
logging.debug('entity_class %s' % entity_class)
logging.info('Reading %s' % csv_path)
headers,rowds = csvfile.make_rowds(fileio.read_csv(csv_path))
logging.info('%s rows' % len(rowds))
# check for modified or uncommitted files in repo
repository = dvcs.repository(cidentifier.path_abs())
logging.debug(repository)
fidentifiers = {
rowd['id']: identifier.Identifier(
id=rowd['id'],
base_path=cidentifier.basepath
)
for rowd in rowds
}
fidentifier_parents = {
fi.id: Importer._fidentifier_parent(fi)
for fi in fidentifiers.itervalues()
}
# eidentifiers, removing duplicates
eidentifiers = list(set([e for e in fidentifier_parents.itervalues()]))
entities = {}
bad_entities = []
for eidentifier in eidentifiers:
if os.path.exists(eidentifier.path_abs()):
entity = eidentifier.object()
entities[eidentifier.id] = entity
else:
if eidentifier.id not in bad_entities:
bad_entities.append(eidentifier.id)
if bad_entities:
for f in bad_entities:
logging.error(' %s missing' % f)
raise Exception('%s entities could not be loaded! - IMPORT CANCELLED!' % len(bad_entities))
# separate into new and existing lists
rowds_new = []
rowds_existing = []
for n,rowd in enumerate(rowds):
if Importer._file_is_new(fidentifiers[rowd['id']]):
rowds_new.append(rowd)
else:
rowds_existing.append(rowd)
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
logging.info('Updating existing files')
start_updates = datetime.now()
git_files = []
updated = []
elapsed_rounds_updates = []
staged = []
obj_metadata = None
for n,rowd in enumerate(rowds_existing):
logging.info('+ %s/%s - %s (%s)' % (n+1, len(rowds), rowd['id'], rowd['basename_orig']))
start_round = datetime.now()
fidentifier = fidentifiers[rowd['id']]
eidentifier = fidentifier_parents[fidentifier.id]
entity = entities[eidentifier.id]
file_ = fidentifier.object()
modified = file_.load_csv(rowd)
# Getting obj_metadata takes about 1sec each time
# TODO caching works as long as all objects have same metadata...
if not obj_metadata:
obj_metadata = models.object_metadata(
fidentifier.fields_module(),
repository.working_dir
)
if dryrun:
pass
elif modified:
logging.debug(' writing %s' % file_.json_path)
file_.write_json(obj_metadata=obj_metadata)
# TODO better to write to collection changelog?
Importer._write_entity_changelog(entity, git_name, git_mail, agent)
# stage
git_files.append(file_.json_path_rel)
git_files.append(entity.changelog_path_rel)
updated.append(file_)
elapsed_round = datetime.now() - start_round
elapsed_rounds_updates.append(elapsed_round)
logging.debug('| %s (%s)' % (fidentifier, elapsed_round))
elapsed_updates = datetime.now() - start_updates
logging.debug('%s updated in %s' % (len(elapsed_rounds_updates), elapsed_updates))
if dryrun:
pass
elif git_files:
logging.info('Staging %s modified files' % len(git_files))
start_stage = datetime.now()
dvcs.stage(repository, git_files)
staged = util.natural_sort(dvcs.list_staged(repository))
for path in staged:
if path in git_files:
logging.debug('+ %s' % path)
else:
logging.debug('| %s' % path)
elapsed_stage = datetime.now() - start_stage
logging.debug('ok (%s)' % elapsed_stage)
logging.debug('%s staged in %s' % (len(staged), elapsed_stage))
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
logging.info('Adding new files')
start_adds = datetime.now()
elapsed_rounds_adds = []
logging.info('Checking source files')
for rowd in rowds_new:
rowd['src_path'] = os.path.join(csv_dir, rowd['basename_orig'])
logging.debug('| %s' % rowd['src_path'])
if not os.path.exists(rowd['src_path']):
raise Exception('Missing file: %s' % rowd['src_path'])
if log_path:
logging.info('addfile logging to %s' % log_path)
for n,rowd in enumerate(rowds_new):
logging.info('+ %s/%s - %s (%s)' % (n+1, len(rowds), rowd['id'], rowd['basename_orig']))
start_round = datetime.now()
fidentifier = fidentifiers[rowd['id']]
eidentifier = fidentifier_parents[fidentifier.id]
entity = entities[eidentifier.id]
logging.debug('| %s' % (entity))
if dryrun:
pass
elif Importer._file_is_new(fidentifier):
# ingest
# TODO make sure this updates entity.files
file_,repo2,log2 = ingest.add_file(
entity,
rowd['src_path'],
fidentifier.parts['role'],
rowd,
git_name, git_mail, agent,
log_path=log_path,
show_staged=False
)
elapsed_round = datetime.now() - start_round
elapsed_rounds_adds.append(elapsed_round)
logging.debug('| %s (%s)' % (file_, elapsed_round))
elapsed_adds = datetime.now() - start_adds
logging.debug('%s added in %s' % (len(elapsed_rounds_adds), elapsed_adds))
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
return git_files
def import_entities(csv_path, cidentifier, vocabs_path, git_name, git_mail, agent, dryrun=False):
"""Adds or updates entities from a CSV file
Running function multiple times with the same CSV file is idempotent.
After the initial pass, files will only be modified if the CSV data
has been updated.
This function writes and stages files but does not commit them!
That is left to the user or to another function.
@param csv_path: Absolute path to CSV data file.
@param cidentifier: Identifier
@param vocabs_path: Absolute path to vocab dir
@param git_name: str
@param git_mail: str
@param agent: str
@param dryrun: boolean
@returns: list of updated entities
"""
logging.info('------------------------------------------------------------------------')
logging.info('batch import entity')
model = 'entity'
repository = dvcs.repository(cidentifier.path_abs())
logging.info(repository)
logging.info('Reading %s' % csv_path)
headers,rowds = csvfile.make_rowds(fileio.read_csv(csv_path))
logging.info('%s rows' % len(rowds))
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
logging.info('Importing')
start_updates = datetime.now()
git_files = []
updated = []
elapsed_rounds = []
obj_metadata = None
if dryrun:
logging.info('Dry run - no modifications')
for n,rowd in enumerate(rowds):
logging.info('%s/%s - %s' % (n+1, len(rowds), rowd['id']))
start_round = datetime.now()
eidentifier = identifier.Identifier(id=rowd['id'], base_path=cidentifier.basepath)
# if there is an existing object it will be loaded
entity = eidentifier.object()
if not entity:
entity = models.Entity.create(eidentifier.path_abs(), eidentifier)
modified = entity.load_csv(rowd)
# Getting obj_metadata takes about 1sec each time
# TODO caching works as long as all objects have same metadata...
if not obj_metadata:
obj_metadata = models.object_metadata(
eidentifier.fields_module(),
repository.working_dir
)
if dryrun:
pass
elif modified:
# write files
if not os.path.exists(entity.path_abs):
os.makedirs(entity.path_abs)
logging.debug(' writing %s' % entity.json_path)
entity.write_json(obj_metadata=obj_metadata)
# TODO better to write to collection changelog?
# TODO write all additions to changelog at one time
Importer._write_entity_changelog(entity, git_name, git_mail, agent)
# stage
git_files.append(entity.json_path_rel)
git_files.append(entity.changelog_path_rel)
updated.append(entity)
elapsed_round = datetime.now() - start_round
elapsed_rounds.append(elapsed_round)
logging.debug('| %s (%s)' % (eidentifier, elapsed_round))
if dryrun:
logging.info('Dry run - no modifications')
elif updated:
logging.info('Staging %s modified files' % len(git_files))
start_stage = datetime.now()
dvcs.stage(repository, git_files)
for path in util.natural_sort(dvcs.list_staged(repository)):
if path in git_files:
logging.debug('+ %s' % path)
else:
logging.debug('| %s' % path)
elapsed_stage = datetime.now() - start_stage
logging.debug('ok (%s)' % elapsed_stage)
elapsed_updates = datetime.now() - start_updates
logging.debug('%s updated in %s' % (len(elapsed_rounds), elapsed_updates))
logging.info('- - - - - - - - - - - - - - - - - - - - - - - -')
return updated
def test_files_import_external_nohashes_rename(tmpdir, collection,
test_csv_dir, test_files_dir):
"""Test importing *external* files with *no* hash cols but binaries present
If file is external, binary is present, and no hash cols, rename binary in place
ddr-testing-123-1-master-684e15e967
ddr-testing-123-2-master-b9773b9aef
"""
print('collection_path %s' % collection.path_abs)
file_csv_path = os.path.join(
test_csv_dir, 'ddrimport-files-import-external-nohashes-rename.csv')
print('file_csv_path %s' % file_csv_path)
rewrite_file_paths(file_csv_path, test_files_dir)
log_path = os.path.join(
test_files_dir, 'ddrimport-files-import-external-nohashes-rename.log')
print('log_path %s' % log_path)
print('test_files_dir %s' % test_files_dir)
for path in os.listdir(test_files_dir):
print(path)
# copy test files so later tests don't crash
# replace basename_orig in CSV with copied file
# and rewrite CSV
headers, rowds, csv_errs = csvfile.make_rowds(
fileio.read_csv(file_csv_path))
renamed_files = []
copied_files = []
ingested_files = []
access_files = []
for rowd in rowds:
print(rowd)
src_file = os.path.join(test_files_dir, rowd['basename_orig'])
path, ext = os.path.splitext(src_file)
dest_file = path + '-rename' + ext
print('shutil.copy(%s, %s)' % (src_file, dest_file))
shutil.copy(src_file, dest_file)
if os.path.exists(dest_file):
renamed_files.append(os.path.basename(dest_file))
else:
print('could not copy')
assert False
rowd['basename_orig'] = dest_file
# figure out new file ID
sha1 = util.file_hash(dest_file, 'sha1')[:10]
idparts = rowd['id'].split('-') + [rowd['role']] + [sha1]
final_file = '-'.join(idparts) + ext
final_access = '-'.join(idparts + ['a.jpg'])
copied_files.append(final_file)
ingested_files.append(final_file)
access_files.append(final_access)
headers, rows = csvfile.make_rows(rowds)
fileio.write_csv(file_csv_path, headers, rows)
out = batch.Importer.import_files(
file_csv_path,
collection.identifier,
VOCABS_URL,
GIT_USER,
GIT_MAIL,
AGENT,
log_path=log_path,
tmp_dir=test_files_dir,
)
# save and commit
repo = dvcs.repository(collection.path_abs)
print('STAGED FILES')
staged_files = sorted([path for path in dvcs.list_staged(repo)])
for path in staged_files:
print(' %s' % path)
# after import_files, we expect to see
offenses = 0
# assert final_file in os.listdir(test_files_dir)
print('test_files_dir')
test_files = [path for path in os.listdir(test_files_dir)]
for path in copied_files:
print(path)
if path not in test_files:
print('RENAMED SRC FILE NOT PRESENT %s' % path)
offenses += 1
# assert files not ingested
# assert no access files created
for path in staged_files:
if os.path.basename(path) in ingested_files:
print('ERROR %s HAS BEEN IMPORTED!!' % path)
offenses += 1
if os.path.basename(path) in access_files:
print('ERROR %s ACCESS FILE GENERATED!!' % path)
offenses += 1
commit = repo.index.commit('test_files_import_external_nohashes_rename')
print('commit %s' % commit)
if offenses:
assert False
# test hashes present
check_file_hashes(collection.path_abs)
# ensure no binaries in .git/objects
print('log_path %s' % log_path)
assert not find_binaries_in_git_objects(repo)
assert not find_missing_annex_binaries(repo)
