Permalink
Commits on Feb 2, 2017
  1. Merge new spelling of "index already exists" so we raise the right ex…

    erikrose committed on GitHub Feb 2, 2017
    …ceptions. Close #195.
Commits on Jun 11, 2015
Commits on Jun 2, 2015
Commits on May 28, 2015
  1. Add HTTPS support. Bump version to 1.3.

    erikrose committed May 28, 2015
Commits on May 22, 2015
  1. Don't crash when the query_params kwarg to send_request is omitted. Fix

    erikrose committed May 22, 2015
    #173.
    
    Keep the null value of the kwarg as None for two reasons: (1) so we don't invalidate the docstring which says we can pass None and (2) because mutables as kwarg default values are asking for trouble.
Commits on Apr 17, 2015
  1. Update AUTHORS file.

    erikrose committed Apr 17, 2015
  2. Honza says they aren't purposely limiting themselves to the lowest co…

    erikrose committed Apr 17, 2015
    …mmon denominator. Close #170.
Commits on Apr 10, 2015
  1. Bring changelog up to date.

    erikrose committed Apr 10, 2015
    I accidentally did a release from my own repo (1.2.1). That’s why 1.2.2 introduces commits that actually predate the ones in 1.2.1.
Commits on Apr 9, 2015
  1. Require elasticsearch lib 1.3 or greater, as that's when it started e…

    erikrose committed Apr 9, 2015
    …xposing `ConnectionTimeout`.
  2. Update ES docs links, now that Elastic has changed domains and reorga…

    erikrose committed Apr 9, 2015
    …nized its docs.
    
    ...and not provided any redirects.
Commits on Mar 31, 2015
  1. Merge pull request #168 from rholder/minor-chunk-documentation-fix

    erikrose committed Mar 31, 2015
    Update client.py docs to reflect how bulk chunking api currently works
Commits on Mar 6, 2015
  1. Make sure the Content-Length header gets set when calling `create_ind…

    erikrose committed Mar 6, 2015
    …ex()` with no explicit `settings` arg. Fix #166. Bump version to 1.2.
Commits on Feb 12, 2015
  1. Rewrite bulk_chunks() procedurally. Address #134.

    erikrose committed Feb 12, 2015
    Now it yields optimal chunks, no longer ever making one over the byte limit unless a single document is over the limit on its own.
    
    I'm ashamed how much simpler it is this way.
  2. Autogenerate PyPI page. Stop repeating things in README.

    erikrose committed Feb 12, 2015
    Swap some header types around to make that possible.
  3. Pretty up the documentation.

    erikrose committed Feb 12, 2015
    * Add new routines.
    * Format the code on the front page more attractively.
    * Show off the interesting bits (bulk()) on the front page.
    * Document JSON encoding customization and HTTP auth.
  4. Introduce new bulk API, supporting all operation types, helping out w…

    erikrose committed Feb 12, 2015
    …ith chunking, detecting per-operation errors, and working across indexes and doc types in a single request. Bump version to 1.1. Close #134, #83, #128.
Commits on Feb 6, 2015
  1. Fix a docs typo.

    erikrose committed Feb 6, 2015
Commits on Jan 26, 2015
  1. This isn't in beta anymore. We're going to strive for backward compat…

    erikrose committed Jan 26, 2015
    …ibility and have deprecation warnings when all else fails.
  2. Remove misleading sentence about retries.

    erikrose committed Jan 26, 2015
    Actually what happens is that your retry preferences apply regardless of whether the failure was due to a connection problem or a timeout, but that's too complicated to get across in a readme.
  3. Patch up the PyPI index page.

    erikrose committed Jan 26, 2015
    * Change repo URL to HTTPS to match the buglink one.
    * Correct version number in elasticsearch-py.rst.
    * Fix changelog extraction to not grab the underline beneath the header. It was showing up as an <hr> (HTML escaped) on PyPI.
Commits on Jan 25, 2015
  1. Merge pull request #161 from HonzaKral/patch-1

    erikrose committed Jan 25, 2015
    Fix broken link to Query DSL
Commits on Jan 23, 2015
  1. Uncap the simplejson version range.

    erikrose committed Jan 23, 2015
    That project doesn't use semantic versioning, and we need >=3.0 for Python 3.3.
  2. Fix test_error_handling in Python 3.4.

    erikrose committed Jan 23, 2015
    urlparse() isn't gullible enough to parse that out-of-range port, so it ended up defaulting to 9200.
  3. Fix a test under Python 3.

    erikrose committed Jan 23, 2015
    We utf-8 encode query param values, so they're binary.