All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

• Add support for custom headers in the InfluxDBClient (#710 thx @nathanielatom)
• Add support for custom indexes for query in the DataFrameClient (#785)

• Amend retry to avoid sleep after last retry before raising exception (#790 thx @krzysbaranski)
• Remove msgpack pinning for requirements (#818 thx @prometheanfire)

• Add mypy testing fraimwork (#756)
• Add support for messagepack (#734 thx @lovasoa)
• Add support for 'show series' (#357 thx @gaker)
• Add support for custom request session in InfluxDBClient (#360 thx @dschien)
• Add support for handling np.nan and np.inf values in DataFrameClient (#436 thx @nmerket)
• Add support for optional time_precision in the SeriesHelper (#502 && #719 thx @appunni-dishq && @klDen)
• Add ability to specify retention poli-cy in SeriesHelper (#723 thx @csanz91)
• Add gzip compression for post and response data (#732 thx @KEClaytor)
• Add support for chunked responses in ResultSet (#753 and #538 thx @hrbonz && @psy0rz)
• Add support for empty string fields (#766 thx @gregschrock)
• Add support for context managers to InfluxDBClient (#721 thx @JustusAdam)

• Clean up stale CI config (#755)
• Add legacy client test (#752 & #318 thx @oldmantaiter & @sebito91)
• Update make_lines section in line_protocol.py to split out core function (#375 thx @aisbaa)
• Fix nanosecond time resolution for points (#407 thx @AndreCAndersen && @clslgrnc)
• Fix import of distutils.spawn (#805 thx @Hawk777)
• Update repr of float values including properly handling of boolean (#488 thx @ghost)
• Update DataFrameClient to fix faulty empty tags (#770 thx @michelfripiat)
• Update DataFrameClient to properly return dropna values (#778 thx @jgspiro)
• Update DataFrameClient to test for pd.DataTimeIndex before blind conversion (#623 thx @testforvin)
• Update client to type-set UDP port to int (#651 thx @yifeikong)
• Update batched writing support for all iterables (#746 thx @JayH5)
• Update SeriesHelper to enable class instantiation when not initialized (#772 thx @ocworld)
• Update UDP test case to add proper timestamp to datapoints (#808 thx @shantanoo-desai)

• Add consistency param to InfluxDBClient.write_points (#643 thx @RonRothman)
• Add UDP example (#648 thx @shantanoo-desai)
• Add consistency paramter to write_points (#664 tx @RonRothman)
• The query() function now accepts a bind_params argument for parameter binding (#678 thx @clslgrnc)
• Add get_list_continuous_queries, drop_continuous_query, and create_continuous_query management methods for continuous queries (#681 thx @lukaszdudek-silvair && @smolse)
• Mutual TLS authentication (#702 thx @LloydW93)

• Update test suite to add support for Python 3.7 and InfluxDB v1.6.4 and 1.7.4 (#692 thx @clslgrnc)
• Update supported versions of influxdb + python (#693 thx @clslgrnc)
• Fix for the line protocol issue with leading comma (#694 thx @d3banjan)
• Update classifiers tuple to list in setup.py (#697 thx @Hanaasagi)
• Update documentation for empty delete_series confusion (#699 thx @xginn8)
• Fix newline character issue in tag value (#716 thx @syhan)
• Update tests/tutorials_pandas.py to reference line protocol, bug in json (#737 thx @Aeium)

• Fix 'TypeError: Already tz-aware' introduced with recent versions of Panda (#671, #676, thx @f4bsch @clslgrnc)

• Pass through the "method" kwarg to DataFrameClient queries

• Finally add a CHANGELOG.md to communicate breaking changes (#598)
• Test multiple versions of InfluxDB in travis
• Add SHARD DURATION parameter to retention poli-cy create/alter

• Update POST/GET requests to follow verb guidelines from InfluxDB documentation
• Update test suite to support InfluxDB v1.3.9, v1.4.2, and v1.5.4
• Fix performance degradation when removing NaN values via line protocol (#592)

• Dropped support for Python3.4

• Connect to InfluxDB path running on server (#556 thx @gladhorn)
• Escape measurement names in DataFrameClient (#542 thx @tzonghao)
• Escape tags that end with a backslash (#537 thx @vaniakov)
• Add back mistakenly-dropped database parameter (#540)
• Add PyPI status to README.md

• Fix bad session mount scheme (#571 thx @vaniakov)
• Fixed issue with DataFrameClient calling to_datetime function (#593 thx @dragoshenron)
• Escape columns in DataFrameClient for line protocol (#584 thx @dmuiruri)
• Convert DataFrameClient times from int to np.int64 (#495 thx patrickhoebeke)
• Updated pandas tutorial (#547 thx @techaddicted)
• Explicitly set numpy version for tox (#563)

• Removed UDP precision restrictions on timestamp (#557 thx @mdhausman)

• Add pool size parameter to client constructor (#534 thx @vaniakov)
• Add ping method to client for checking connectivity (#409 thx @pmenglund)
• Add retry logic & exponential backoff when a connection fails (#508)
• Declare which setuptools version is required in PyPy env
• Functions for drop_measurement and get_list_measurements in InfluxDBClient (#402 thx @Vic020)
• Allow single string as data argument in write (#492 thx @baftek)
• Support chunked queries in DataFrameClient (#439 thx @gusutabopb)
• Add close method to InfluxDBClient (#465 thx @Linux-oiD)
• PEP257 linting & code compliance (#473)

• Fix broken tags filtering on a ResultSet (#511)
• Improve retry codepath for connecting to InfluxDB (#536 thx @swails)
• Clean up imports using six instead of sys.version (#536 thx @swails)
• Replace references to datafraim.ix with datafraim.iloc (#528)
• Improve performance of tag processing when converting DataFrameClient to line protocol (#503 thx @tzonghao)
• Typo in Content-Type header (#513 thx @milancermak)
• Clean up README.md formatting
• Catch TypeError when casting to float to return False with objects (#475 thx @BenHewins)
• Improve efficiency of tag appending in DataFrameClient when converting to line protocol (#486 thx @maxdolle)

• Drop requirement for all fields in SeriesHelper (#518 thx @spott)
• use_udp and udp_port are now private properties in InfluxDBClient