def __init__(self, plugin_path=None):
self.files_manager = FilesManager()
self.plugin_path = plugin_path
class PluginLoaderBase(object):
"""Plugin loader.
Loader deals with the file structure providing ability to load, combine
and form the data tree from the plugin directory.
If loader fails it raising exception with the report attached.
"""
_metadata_path = "metadata.yaml"
_path_suffix = "_path"
_dont_resolve_path_keys = {'repository_path', 'deployment_scripts_path'}
def __init__(self, plugin_path=None):
self.files_manager = FilesManager()
self.plugin_path = plugin_path
paths_to_fields = {}
def _get_absolute_path(self, path):
"""Get absolute path from the relative to the plugins folder.
:param path: relative path
:type path: str
:return: path string
:rtype: str
"""
return os.path.join(self.plugin_path, path)
@property
def _root_metadata_path(self):
"""Where is the root plugin data file located."""
return self._get_absolute_path(self._metadata_path)
def _recursive_process_paths(self, data, report):
"""Recursively processed nested list/dict.
:param data: data
:type data: iterable
:param report: report node
:type report: utils.ReportNode
:returns: data
:rtype: list|dict
"""
if isinstance(data, dict):
new_data = {}
for key in tuple(data):
value = data[key]
# if we have key with path we could do 3 things:
#
# * if it is pointing to directory, check dir existence and
# leave path intact
#
# * if it is a `glob` compatible mask, iterate over files
# that are matched this mask and compatible with
# FileManager then merge this files data if they have
# list or dict as common data root.
# Then remove _path suffix from key.
#
# * if it is file compatible with FileManager, read this
# file and remove _path suffix from key.
if key.endswith(self._path_suffix) \
and isinstance(value, six.string_types):
if os.path.isdir(self._get_absolute_path(value)):
report.info(u"{} is valid directory".format(
value))
# leave directories as is
new_data[key] = value
elif key in self._dont_resolve_path_keys:
report.info(u"{}:{} was not checked".format(
key, value))
new_data[key] = value
else:
cleaned_key = key[:- len(self._path_suffix)]
try:
loaded_data = self.files_manager.load(
self._get_absolute_path(value)
)
new_data[cleaned_key] = loaded_data
except Exception as exc:
path_node = utils.ReportNode(data[key])
report.add_nodes(path_node.error(exc))
# keep path as is
new_data[key] = value
else:
new_data[key] = self._recursive_process_paths(
data[key], report)
elif isinstance(data, list):
new_data = [
self._recursive_process_paths(record, report)
for record in data
]
else:
new_data = data
return new_data
def _load_root_metadata_file(self):
"""Get plugin root data (usually, it's metadata.yaml).
:return: data
:rtype: list|dict
"""
report = utils.ReportNode(u"Loading root metadata file:{}".format(
self._root_metadata_path
))
# todo(ikutukov): current loading schema and testing relies on case
# when no metadata.yaml file is possible. So we are skipping all
# exeptions.
try:
data = self.files_manager.load(self._root_metadata_path)
except Exception as exc:
report.warning(exc)
return {}, report
data = self._recursive_process_paths(data, report)
return data, report
def load(self, plugin_path=None):
"""Loads data from the given plugin path and producing data tree.
:param plugin_path: plugin root path
:param plugin_path: str|basestring|None
:return: data tree starting from the data in root metadata file
:rtype: tuple(dict, utils.ReportNode)
"""
plugin_path = plugin_path or self.plugin_path
report = utils.ReportNode(
u"File structure validation: {}".format(plugin_path))
data, root_report = self._load_root_metadata_file()
report.add_nodes(root_report)
# load files with fixed location
for key, file_path in six.iteritems(self.paths_to_fields):
file_report = utils.ReportNode(file_path)
try:
data[key] = self.files_manager.load(
self._get_absolute_path(file_path)
)
except errors.NoPluginFileFound as exc:
data[key] = None
file_report.warning(exc)
except Exception as exc:
file_report.error(exc)
finally:
report.add_nodes(file_report)
if report.is_failed():
raise errors.ParseError(report.render())
return data, report
def __init__(self, plugin_path=None):
self.files_manager = FilesManager()
self.plugin_path = plugin_path
class PluginLoaderBase(object):
"""Plugin loader.
Loader deals with the file structure providing ability to load, combine
and form the data tree from the plugin directory.
If loader fails it raising exception with the report attached.
"""
_metadata_path = "metadata.yaml"
_path_suffix = "_path"
_dont_resolve_path_keys = {'repository_path', 'deployment_scripts_path'}
def __init__(self, plugin_path=None):
self.files_manager = FilesManager()
self.plugin_path = plugin_path
paths_to_fields = {}
def _get_absolute_path(self, path):
"""Get absolute path from the relative to the plugins folder.
:param path: relative path
:type path: str
:return: path string
:rtype: str
"""
return os.path.join(self.plugin_path, path)
@property
def _root_metadata_path(self):
"""Where is the root plugin data file located."""
return self._get_absolute_path(self._metadata_path)
def _recursive_process_paths(self, data, report):
"""Recursively processed nested list/dict.
:param data: data
:type data: iterable
:param report: report node
:type report: utils.ReportNode
:returns: data
:rtype: list|dict
"""
if isinstance(data, dict):
new_data = {}
for key in tuple(data):
value = data[key]
# if we have key with path we could do 3 things:
#
# * if it is pointing to directory, check dir existence and
# leave path intact
#
# * if it is a `glob` compatible mask, iterate over files
# that are matched this mask and compatible with
# FileManager then merge this files data if they have
# list or dict as common data root.
# Then remove _path suffix from key.
#
# * if it is file compatible with FileManager, read this
# file and remove _path suffix from key.
if key.endswith(self._path_suffix) \
and isinstance(value, six.string_types):
if os.path.isdir(self._get_absolute_path(value)):
report.info(u"{} is valid directory".format(value))
# leave directories as is
new_data[key] = value
elif key in self._dont_resolve_path_keys:
report.info(u"{}:{} was not checked".format(
key, value))
new_data[key] = value
else:
cleaned_key = key[:-len(self._path_suffix)]
try:
loaded_data = self.files_manager.load(
self._get_absolute_path(value))
new_data[cleaned_key] = loaded_data
except Exception as exc:
path_node = utils.ReportNode(data[key])
report.add_nodes(path_node.error(exc))
# keep path as is
new_data[key] = value
else:
new_data[key] = self._recursive_process_paths(
data[key], report)
elif isinstance(data, list):
new_data = [
self._recursive_process_paths(record, report)
for record in data
]
else:
new_data = data
return new_data
def _load_root_metadata_file(self):
"""Get plugin root data (usually, it's metadata.yaml).
:return: data
:rtype: list|dict
"""
report = utils.ReportNode(u"Loading root metadata file:{}".format(
self._root_metadata_path))
# todo(ikutukov): current loading schema and testing relies on case
# when no metadata.yaml file is possible. So we are skipping all
# exeptions.
try:
data = self.files_manager.load(self._root_metadata_path)
except Exception as exc:
report.warning(exc)
return {}, report
data = self._recursive_process_paths(data, report)
return data, report
def load(self, plugin_path=None):
"""Loads data from the given plugin path and producing data tree.
:param plugin_path: plugin root path
:param plugin_path: str|basestring|None
:return: data tree starting from the data in root metadata file
:rtype: tuple(dict, utils.ReportNode)
"""
plugin_path = plugin_path or self.plugin_path
report = utils.ReportNode(
u"File structure validation: {}".format(plugin_path))
data, root_report = self._load_root_metadata_file()
report.add_nodes(root_report)
# load files with fixed location
for key, file_path in six.iteritems(self.paths_to_fields):
file_report = utils.ReportNode(file_path)
try:
data[key] = self.files_manager.load(
self._get_absolute_path(file_path))
except errors.NoPluginFileFound as exc:
data[key] = None
file_report.warning(exc)
except Exception as exc:
file_report.error(exc)
finally:
report.add_nodes(file_report)
if report.is_failed():
raise errors.ParseError(report.render())
return data, report
def __init__(self, prefix, output_size_mb, dbx):
#initialize dropbox connector
self.dbx = dbx
self.prefix = prefix
self.files_manager = FilesManager(output_size_mb)
class Processor:
def __init__(self, prefix, output_size_mb, dbx):
#initialize dropbox connector
self.dbx = dbx
self.prefix = prefix
self.files_manager = FilesManager(output_size_mb)
def set_log(self, log):
self.log = log
self.files_manager.set_log(log)
def TL_data(self, data_folder_dropbox, dropbox_folder_upload, data_folder,
result_folder, sufix, category):
filenames = self.dbx.list_files(data_folder_dropbox + category)
final_df = pd.DataFrame()
index = 0
dropbox_path = dropbox_folder_upload + category
if (not self.dbx.folder_exists(dropbox_path)):
self.log.info('Create folder ' + dropbox_path)
self.dbx.create_folder(dropbox_path)
#for each file, append until reach threashold
for filename in filenames:
path = data_folder_dropbox + category + '/' + filename
local_path = data_folder + filename
self.dbx.download_file(path, local_path)
self.log.info('Processing file:%s', local_path)
try:
df = pd.read_csv(local_path, dtype={'ticker': str})
if (df.shape[0] == 0):
os.remove(local_path)
continue
except pd.io.common.EmptyDataError:
os.remove(local_path)
continue
ticker = df['ticker'][0]
df.rename(index=str, columns={ticker: 'count'}, inplace=True)
if (self.files_manager.check_chunks(final_df, df)):
self.log.info('Uploading chunk:%d', index)
#sort values by date before saving the file
final_df.sort_values(by=['date'],
inplace=True,
ascending=False)
index = self.files_manager.save_data(
result_folder, self.prefix, sufix,
dropbox_folder_upload + category + '/', final_df, index,
self.dbx)
final_df = df
else:
final_df = final_df.append(df)
os.remove(local_path)
#save last chunck
if (final_df.shape[0] > 0):
self.log.info('Saving last chunck')
final_df.sort_values(by=['date'], inplace=True, ascending=False)
self.files_manager.save_data(
result_folder, self.prefix, sufix,
dropbox_folder_upload + category + '/', final_df, index,
self.dbx)
