def index_location(args, config):
path = config['location']
add_site_libs(path)
# This whole routine is probably being too careful
# Explicit is better than implicit, though!
index_processor_helper = IndexHelper()
index_processor_helper.configure(config)
# the IndexHelper singleton can be used in processors that
# need to talk to elasticsearch
settings_path = os.path.join(path, '_settings/settings.json')
default_mapping_path = os.path.join(path, '_defaults/mappings.json')
processors_path = os.path.join(path, '_settings/processors.json')
es = Elasticsearch(config["elasticsearch"])
index_name = config["index"]
if es.indices.exists(index_name):
es.indices.delete(index_name)
if os.path.exists(settings_path):
es.indices.create(index=index_name, body=file(settings_path).read())
else:
es.indices.create(index=index_name)
processors = []
processor_settings = read_json_file(processors_path)
if processor_settings:
configured_processors = [
ContentProcessor(name, **details)
for name, details in processor_settings.iteritems()
]
processors += configured_processors
glob_pattern = "".join([os.path.normpath(path), "/_*/"])
underscored = glob.glob(glob_pattern)
ignore_dirs = [os.path.join(path, d) for d in DO_NOT_INDEX]
filesystem_candidates = [u for u in underscored if u not in ignore_dirs]
for f in filesystem_candidates:
# TODO: don't create processors for directories that
# have a configured processor
processor_name_starts = f[0:-1].rfind('/') + 2
processor_name = f[processor_name_starts:-1]
processor_args = dict(directory=f,
site_root=path,
processor="sheer.processors.filesystem")
processors.append(ContentProcessor(processor_name, **processor_args))
# Load default mapping (or not)
if os.path.exists(default_mapping_path):
try:
default_mapping = read_json_file(default_mapping_path)
except ValueError:
sys.exit("default mapping present, but is not valid JSON")
else:
default_mapping = {}
for processor in processors:
print "creating mapping for %s (%s)" % (processor.name,
processor.processor_name)
es.indices.put_mapping(
index=index_name,
doc_type=processor.name,
body={processor.name: processor.mapping(default_mapping)})
i = -1
for i, document in enumerate(processor.documents()):
es.create(index=index_name,
doc_type=processor.name,
id=document['_id'],
body=document)
sys.stdout.write("indexed %s %s \r" % (i + 1, processor.name))
sys.stdout.flush()
sys.stdout.write("indexed %s %s \n" % (i + 1, processor.name))
def index_location(args, config):
path = config['location']
add_site_libs(path)
# This whole routine is probably being too careful
# Explicit is better than implicit, though!
index_processor_helper = IndexHelper()
index_processor_helper.configure(config)
# the IndexHelper singleton can be used in processors that
# need to talk to elasticsearch
settings_path = os.path.join(path, '_settings/settings.json')
processors_path = os.path.join(path, '_settings/processors.json')
es = Elasticsearch(config["elasticsearch"])
index_name = config["index"]
# If we're given args.reindex and NOT given a list of processors to reindex,
# we're expected to reindex everything. Delete the existing index.
if not args.processors and args.reindex and es.indices.exists(index_name):
print "reindexing %s" % index_name
es.indices.delete(index_name)
# If the index doesn't exist, create it.
if not es.indices.exists(index_name):
if os.path.exists(settings_path):
es.indices.create(index=index_name, body=file(settings_path).read())
else:
es.indices.create(index=index_name)
processors = []
processor_settings = read_json_file(processors_path)
if processor_settings:
configured_processors = [ContentProcessor(name, **details)
for name, details
in processor_settings.iteritems()]
processors += configured_processors
glob_pattern = "".join([os.path.normpath(path), "/_*/"])
underscored = glob.glob(glob_pattern)
ignore_dirs = [os.path.join(path, d) for d in DO_NOT_INDEX]
filesystem_candidates = [u for u in underscored if u not in ignore_dirs]
for f in filesystem_candidates:
# TODO: don't create processors for directories that
# have a configured processor
processor_name_starts = f[0:-1].rfind('/') + 2
processor_name = f[processor_name_starts:-1]
processor_args = dict(directory=f,
site_root=path,
processor="sheer.processors.filesystem")
processors.append(ContentProcessor(processor_name, **processor_args))
# If any specific content processors were selected, we run them. Otherwise
# we run all of them.
selected_processors = processors
if args.processors and len(args.processors) > 0:
selected_processors = [p for p in processors if p.name in args.processors]
failed_processors = []
for processor in selected_processors:
index_sucess = index_processor(es,
index_name,
processor,
reindex=args.reindex)
if not index_sucess:
failed_processors.append(processor.name)
# Exit with an error code != 0 if there were any issues with indexing
if failed_processors:
sys.exit("Indexing the following processor(s) failed: {}".format(
", ".join(failed_processors)))