aboutsummaryrefslogtreecommitdiffstats
path: root/third_party/influxdb
diff options
context:
space:
mode:
Diffstat (limited to 'third_party/influxdb')
-rw-r--r--third_party/influxdb/__init__.py0
-rw-r--r--third_party/influxdb/influxdb_line_protocol.py137
2 files changed, 137 insertions, 0 deletions
diff --git a/third_party/influxdb/__init__.py b/third_party/influxdb/__init__.py
new file mode 100644
index 000000000..e69de29bb
--- /dev/null
+++ b/third_party/influxdb/__init__.py
diff --git a/third_party/influxdb/influxdb_line_protocol.py b/third_party/influxdb/influxdb_line_protocol.py
new file mode 100644
index 000000000..eee982163
--- /dev/null
+++ b/third_party/influxdb/influxdb_line_protocol.py
@@ -0,0 +1,137 @@
+# The MIT License (MIT)
+
+# Copyright (c) 2013 InfluxDB
+
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in
+# the Software without restriction, including without limitation the rights to
+# use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
+# of the Software, and to permit persons to whom the Software is furnished to
+# do so, subject to the following conditions:
+
+# The above copyright notice and this permission notice shall be included in
+# all copies or substantial portions of the Software.
+
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+
+# yardstick comment: this file is a modified copy of
+# influxdb-python/influxdb/line_protocol.py
+
+from __future__ import unicode_literals
+from copy import copy
+
+from six import binary_type, text_type, integer_types
+
+
+def _escape_tag(tag):
+ tag = _get_unicode(tag, force=True)
+ return tag.replace(
+ "\\", "\\\\"
+ ).replace(
+ " ", "\\ "
+ ).replace(
+ ",", "\\,"
+ ).replace(
+ "=", "\\="
+ )
+
+
+def _escape_value(value):
+ value = _get_unicode(value)
+ if isinstance(value, text_type) and value != '':
+ return "\"{}\"".format(
+ value.replace(
+ "\"", "\\\""
+ ).replace(
+ "\n", "\\n"
+ )
+ )
+ elif isinstance(value, integer_types) and not isinstance(value, bool):
+ return str(value) + 'i'
+ else:
+ return str(value)
+
+
+def _get_unicode(data, force=False):
+ """
+ Try to return a text aka unicode object from the given data.
+ """
+ if isinstance(data, binary_type):
+ return data.decode('utf-8')
+ elif data is None:
+ return ''
+ elif force:
+ return str(data)
+ else:
+ return data
+
+
+def make_lines(data):
+ """
+ Extracts the points from the given dict and returns a Unicode string
+ matching the line protocol introduced in InfluxDB 0.9.0.
+
+ line protocol format:
+ <measurement>[,<tag-key>=<tag-value>...] <field-key>=<field-value>\
+ [,<field2-key>=<field2-value>...] [unix-nano-timestamp]
+
+ Ref:
+ https://influxdb.com/docs/v0.9/write_protocols/write_syntax.html
+ https://influxdb.com/docs/v0.9/write_protocols/line.html
+ """
+ lines = []
+ static_tags = data.get('tags', None)
+ for point in data['points']:
+ elements = []
+
+ # add measurement name
+ measurement = _escape_tag(_get_unicode(
+ point.get('measurement', data.get('measurement'))
+ ))
+ key_values = [measurement]
+
+ # add tags
+ if static_tags is None:
+ tags = point.get('tags', {})
+ else:
+ tags = copy(static_tags)
+ tags.update(point.get('tags', {}))
+
+ # tags should be sorted client-side to take load off server
+ for tag_key in sorted(tags.keys()):
+ key = _escape_tag(tag_key)
+ value = _escape_tag(tags[tag_key])
+
+ if key != '' and value != '':
+ key_values.append("{key}={value}".format(key=key, value=value))
+ key_values = ','.join(key_values)
+ elements.append(key_values)
+
+ # add fields
+ field_values = []
+ for field_key in sorted(point['fields'].keys()):
+ key = _escape_tag(field_key)
+ value = _escape_value(point['fields'][field_key])
+ if key != '' and value != '':
+ field_values.append("{key}={value}".format(
+ key=key,
+ value=value
+ ))
+ field_values = ','.join(field_values)
+ elements.append(field_values)
+
+ # add timestamp
+ if 'time' in point:
+ elements.append(point['time'])
+
+ line = ' '.join(elements)
+ lines.append(line)
+ lines = '\n'.join(lines)
+ return lines + '\n'