| 0b436e1 | Dave Eargle | 09 August 2017, 18:21:07 UTC | bump version | 09 August 2017, 18:21:07 UTC |
| e88f265 | Dave Eargle | 09 August 2017, 18:20:16 UTC | Merge pull request #281 from NYUCCL/db-mode-column add db mode column | 09 August 2017, 18:20:16 UTC |
| 6b55577 | Dave Eargle | 02 August 2017, 22:45:04 UTC | bumping to version 2.2.2 | 02 August 2017, 22:45:04 UTC |
| b7bee89 | Dave Eargle | 26 July 2017, 18:43:35 UTC | specify heroku python runtime | 26 July 2017, 18:43:35 UTC |
| 7100397 | Todd Gureckis | 27 June 2017, 22:04:38 UTC | closes #280 documentation typo | 27 June 2017, 22:04:38 UTC |
| f6e6a55 | Dave Eargle | 01 June 2017, 00:54:57 UTC | removing inappropriate call to 'finish' | 01 June 2017, 00:54:57 UTC |
| 309af67 | Dave Eargle | 31 May 2017, 20:12:59 UTC | moving InvalidUsage out of experiment.py, adding to custom.py default imports | 31 May 2017, 20:12:59 UTC |
| 4ef899e | Dave Eargle | 31 May 2017, 18:43:41 UTC | pull heroku database url from env | 31 May 2017, 18:43:41 UTC |
| 82e2373 | Dave Eargle | 31 May 2017, 18:11:26 UTC | adding an InvalidUsage exception handler for use with API routes | 31 May 2017, 18:11:26 UTC |
| 2009ced | David Eargle | 14 May 2017, 03:50:09 UTC | Merge branch 'master' of github.com:NYUCCL/psiTurk | 14 May 2017, 03:50:09 UTC |
| d862fb4 | Dave Eargle | 14 May 2017, 03:41:49 UTC | Merge pull request #278 from ymollard/master Exit the worker command when not connected | 14 May 2017, 03:41:49 UTC |
| dfaebcb | Dave Eargle | 14 May 2017, 03:38:45 UTC | Merge pull request #279 from NYUCCL/bump_flask_werkzeug bump Flask and Werkzeug versions | 14 May 2017, 03:38:45 UTC |
| 0c24c45 | David Eargle | 14 May 2017, 02:51:40 UTC | Retry submission to mturk if failed previously `or status == SUBMITTED` because we suspect that sometimes the post to mturk fails after we've set status to SUBMITTED, so really they have not successfully submitted. This gives another chance for the submit to work when not using the psiturk ad server. Note that, unfortunately, we cannot post to turkSubmitTo on behalf of the turkers in order that we might detect any error response. MTurk requires that the post originate from the turker's (authenticated) browser -- and that post cannot happen over AJAX, either :( A fix for the psiturk ad server will be harder because the ad server does not understand the notion of 'already_in_db', which is crucial for this approach to work. `already_in_db == false` + `status == submitted` ensures that we're looking at a worker that hasn't already completed an assignment for us. The ad server assumes that anyone with `status > completed` has already successfully submitted. `/check_worker_status` will need an overhaul for the ad server to be able to retry posts. | 14 May 2017, 02:59:43 UTC |
| eeb56fc | David Eargle | 14 May 2017, 02:40:22 UTC | mode-ify /complete route The /complete route now relies on the mode request arg being passed, so that mode=lab can be used without requiring that the hitId and assignmentId be prefixed with the 'debug' string. | 14 May 2017, 02:40:22 UTC |
| 155e5a8 | Alex Rich | 13 May 2017, 19:44:30 UTC | bump Flask and Werkzeug versions Flask 0.12.1 allows range requests, allowing proper streaming and seeking for videos. | 13 May 2017, 19:44:30 UTC |
| ee06b90 | Yoan Mollard | 24 April 2017, 18:11:15 UTC | Exit the worker command when not connected | 24 April 2017, 18:11:15 UTC |
| 45d3d96 | Dave Eargle | 28 February 2017, 23:04:50 UTC | bumping requests closes #268 | 28 February 2017, 23:04:50 UTC |
| e755c65 | Dave Eargle | 28 February 2017, 03:02:43 UTC | bump version | 28 February 2017, 03:02:43 UTC |
| b86ef13 | Dave Eargle | 28 February 2017, 01:27:46 UTC | cast adserver_revproxy_port else to str Fixes a bug where if adserver_revproxy_port is not set, there is an attempt to concat an int with a str, which caused an explosion. Thanks @braingineer for noticing | 28 February 2017, 01:27:46 UTC |
| 07665d8 | Dave Eargle | 27 February 2017, 21:24:23 UTC | adding heroku files to example | 27 February 2017, 21:24:23 UTC |
| b757fec | Dave Eargle | 27 February 2017, 20:59:47 UTC | dynamically set port if on heroku | 27 February 2017, 20:59:47 UTC |
| a8afaeb | Dave Eargle | 27 February 2017, 20:40:05 UTC | fixing logging on heroku | 27 February 2017, 20:40:05 UTC |
| f0b6d8f | Dave Eargle | 27 February 2017, 20:33:30 UTC | improving loggin on heroku | 27 February 2017, 20:33:30 UTC |
| d6a6789 | Dave Eargle | 27 February 2017, 20:10:56 UTC | overriding global configs with env vars if set | 27 February 2017, 20:10:56 UTC |
| 8bb0740 | Dave Eargle | 27 February 2017, 19:46:06 UTC | removing unused config parsers removed both localParser and globalParser. also removed write and set() functions. Assuming they're holdovers from the web gui psiturk version (pre-shell days). | 27 February 2017, 19:46:06 UTC |
| a8345d1 | Dave Eargle | 15 February 2017, 15:23:26 UTC | bumping version | 15 February 2017, 15:23:26 UTC |
| 0d61195 | Dave Eargle | 15 February 2017, 04:05:55 UTC | Merge branch 'allow-repeats-to-adserver' | 15 February 2017, 04:05:55 UTC |
| 9ae3b32 | Dave Eargle | 15 February 2017, 04:00:41 UTC | removing warning about ad server not supporting allow_repeats | 15 February 2017, 04:00:41 UTC |
| 55206bf | Dave Eargle | 15 February 2017, 03:55:29 UTC | Merge branch '263-qualifications-fix' | 15 February 2017, 03:55:29 UTC |
| e8dcd35 | Dave Eargle | 15 February 2017, 03:54:55 UTC | removing ignored common parameter values during hit creation | 15 February 2017, 03:54:55 UTC |
| 4d6c0a8 | chepner3 | 14 February 2017, 22:54:35 UTC | Added qualification requirements to HIT type (Fixes #263) | 14 February 2017, 22:54:35 UTC |
| 3d77ccf | Todd Gureckis | 09 February 2017, 18:44:50 UTC | Issue #90: proposal for dealing with repeats if allow_repeats = False after a period where it is True then use the highest status associated with that worker as the exclusion criteria. | 09 February 2017, 18:44:50 UTC |
| dd89fe8 | Todd Gureckis | 09 February 2017, 17:01:11 UTC | always check worker status w.r.t. current assignmentId | 09 February 2017, 17:01:11 UTC |
| 0027258 | Dave Eargle | 08 February 2017, 21:55:03 UTC | passing allow_repeats to the adserver | 08 February 2017, 21:55:03 UTC |
| d00b571 | Dave Eargle | 08 February 2017, 21:14:04 UTC | adding a missing chdir in an overriding teardown | 08 February 2017, 21:14:04 UTC |
| f66891a | Dave Eargle | 08 February 2017, 21:13:29 UTC | adding a warning about allow_repeats not yet being implemented on the psiturk ad server | 08 February 2017, 21:13:29 UTC |
| 9fe45d3 | Dave Eargle | 08 February 2017, 21:13:02 UTC | removing reference to a gist that no longer exists | 08 February 2017, 21:13:02 UTC |
| 0898a88 | Dave Eargle | 08 February 2017, 21:07:29 UTC | Merge branch 'jhamrick-allow-repeats' Note that this is not yet implemented on the psiturk ad server. So you will need to host your own ad if you want to use this feature. | 08 February 2017, 21:07:29 UTC |
| 3095462 | Dave Eargle | 07 February 2017, 01:00:35 UTC | removing warning message from docs for num_conds see bugfix in #257 | 07 February 2017, 01:00:35 UTC |
| 07a7e9a | Dave Eargle | 07 February 2017, 00:55:39 UTC | Merge pull request #257 from ariekahn/fix/condupdate always use current cond/counterbalance values from config | 07 February 2017, 00:55:39 UTC |
| e526c5b | Ari Kahn | 06 February 2017, 23:12:00 UTC | always use current cond/counterbalance values from config | 06 February 2017, 23:12:00 UTC |
| f66f3a6 | Dave Eargle | 06 February 2017, 19:10:03 UTC | point ad server to reverse proxy This adds two new config params that allow a user to tell the psiturk ad server to direct traffic to a reverse proxy, such as nginx or apache, instead of directly to the psiturk gunicorn server. If the `adserver_revproxy_host` param is present (not commented out), then its value will take precedence over all others when a debugging url is requested (e.g., `debug` or `debug -p`). See also discussion in #246 | 06 February 2017, 19:15:30 UTC |
| db7fb72 | Dave Eargle | 03 February 2017, 21:12:40 UTC | Merge pull request #256 from jacob-lee/master Bug Fixes and Improvements to Hit Creation | 03 February 2017, 21:12:40 UTC |
| 14fe4c3 | Jacob Lee | 30 January 2017, 17:17:39 UTC | Committing for testing. Bug Fixes and Improvements to Hit Creation The following bugs and UX issues were noticed when creating a HIT in the command line tool: 1. Partial command input raises uncaught exception Example Input: hit create <numWorkers> Now it queries for the remaining arguments. 2. The regex to validate the reward per HIT fails on inputs like: -10.00 3J55 J55 New stricter regex used, along with cast to float. 3. Uncaught exception when duration can not be cast as a float, e.g. when receiving a non-numeric string as float. Note, the arguments passed to amt_services methods have been cast to their numeric equivalents. My tests, and the documentation for boto.mturk and amazon's amt API suggest this should be just fine. | 30 January 2017, 18:18:03 UTC |
| 8464eaf | Dave Eargle | 25 January 2017, 03:29:56 UTC | bumping version | 25 January 2017, 03:29:56 UTC |
| 4fc0164 | Dave Eargle | 24 January 2017, 01:14:41 UTC | Merge pull request #255 from NYUCCL/173-import-gnureadline-if-available import gnureadline if available | 24 January 2017, 01:14:41 UTC |
| 427ead5 | Dave Eargle | 24 January 2017, 01:10:22 UTC | I don't pretend to be expert in mac enviornments, but explicitly importing gnureadline fixes broken things see #173 | 24 January 2017, 01:10:22 UTC |
| 3273512 | Dave Eargle | 05 January 2017, 02:28:09 UTC | Merge pull request #253 from jacob-lee/master The methods get_trial_data, get_question_data and get_eventdata were not catching TypeError and ValueError exceptions as intended, causing data download in the psiturk shell to fail. The list of exception classes needs to be enclosed in parentheses. | 05 January 2017, 02:28:09 UTC |
| a391711 | Jacob Lee | 30 December 2016, 22:25:45 UTC | Bug Fix The methods get_trial_data, get_question_data and get_eventdata were not catching TypeError and ValueError exceptions as intended due to a lack of parentheses surrounding the list of exceptions in the except clauses. | 30 December 2016, 22:25:45 UTC |
| aeda804 | Dave Eargle | 04 October 2016, 20:47:19 UTC | Merge branch 'master' into db-mode-column | 04 October 2016, 20:47:19 UTC |
| 388b22c | Dave Eargle | 04 October 2016, 20:46:40 UTC | adding a `pub` alias for the `ad` route, to get around ad blockers. will need more work later to interface with the psiturk ad server | 04 October 2016, 20:46:40 UTC |
| f6a90fa | Dave Eargle | 04 October 2016, 18:22:06 UTC | adding a new column, to make data analysis easier to filter out sandboxed and debug hits | 04 October 2016, 18:22:06 UTC |
| 30d8461 | Dave Eargle | 06 September 2016, 04:05:28 UTC | Merge branch 'allow-repeats' of https://github.com/jhamrick/psiTurk into jhamrick-allow-repeats Also fix a bug in the test_psiturk that was not finding an error string because of html formatting difference | 06 September 2016, 04:05:28 UTC |
| 28593ff | Dave Eargle | 06 September 2016, 03:28:24 UTC | allowing psutil to be higher to fix an anaconda issue | 06 September 2016, 03:28:24 UTC |
| 22363f9 | Dave Eargle | 06 September 2016, 03:26:40 UTC | Merge pull request #249 from pmlandwehr/patch-1 Adding a reference to the LICENSE | 06 September 2016, 03:26:40 UTC |
| 8f534a0 | Dave Eargle | 23 August 2016, 20:32:01 UTC | bumpity-bump-bump | 23 August 2016, 20:32:01 UTC |
| c375bfc | Dave Eargle | 23 August 2016, 20:24:34 UTC | bumping again for travis' sake | 23 August 2016, 20:24:34 UTC |
| 9bc9d85 | Dave Eargle | 23 August 2016, 20:18:23 UTC | Update version.py | 23 August 2016, 20:18:23 UTC |
| 1f7fbac | Dave Eargle | 23 August 2016, 19:31:16 UTC | updating travis yml config for pypi deploy hook | 23 August 2016, 19:31:16 UTC |
| 0723cbc | Peter M. Landwehr | 17 August 2016, 18:00:22 UTC | Adding a reference to the LICENSE I've gone ahead and made a `conda` package out of this repo for distribution via the community-curated `conda-forge` channel. (If y'all want to help maintain it, that would be cool, but is unrelated.) We've started included links to license files in the conda builds themselves, but doing so requires that the license be in the manifest itself. If this pull gets incorporated into the next pypi version, that would let us include said link. | 17 August 2016, 18:00:22 UTC |
| 44e3836 | Dave Eargle | 15 August 2016, 19:40:52 UTC | Merge pull request #248 from NYUCCL/execute_one_cmd making it possible to execute just one command instead of having to r… | 15 August 2016, 19:40:52 UTC |
| 8182f74 | Dave Eargle | 03 August 2016, 06:38:58 UTC | making it possible to execute just one command instead of having to read in a script. this will facilitate batch-creating hits via `hit create`. | 03 August 2016, 06:39:22 UTC |
| 75c078e | Todd Gureckis | 01 August 2016, 03:55:19 UTC | due to pypi problem just stepping minor version forward by one (sorry y'all!) | 01 August 2016, 03:55:19 UTC |
| f9d5ce2 | Todd Gureckis | 29 July 2016, 19:43:09 UTC | minor fix for pypi release | 29 July 2016, 19:43:09 UTC |
| 3827d9e | Todd Gureckis | 28 July 2016, 18:36:18 UTC | preparing 2.1.3 release | 28 July 2016, 18:36:18 UTC |
| 792b8ab | Dave Eargle | 14 July 2016, 21:20:21 UTC | nuking container-ad strings (#245) see also 1f9d16e | 14 July 2016, 21:20:21 UTC |
| d7d9eba | Dave Eargle | 13 July 2016, 17:53:55 UTC | Update thanks-mturksubmit.html making the error message more user-friendly. | 13 July 2016, 17:53:55 UTC |
| 178a21d | Dave Eargle | 12 July 2016, 01:16:38 UTC | Use own ad server (#226) * documenting the tab-complete functions * adding functionality for gunicorn ssl and for specifying one's own ad_location (instead of using the psiturk location) * printing the ad url and the mturk url, for both psiturk and non-psiturk ad_locations * no need to ask them twice if they're using an external server... * functionizes generate_hit_config and also create_psiturk_ad to make hit_create less bulky * Adding some documentation for the new [server parameter] and [shell parameter] SSL- and ad-server-related experimental features. Also I'm removing most of the inline documentation from the default local config.txt file. It seemed like it might have been overwhelming to newcomers. I moved it into the docs instead. * added important note to docs emphasizing that if the psiturk server is ssl-enabled, a proxy server must be used in front of it in order to serve static content. * warn before reloading on submit page if it hasn't finished submitting to mturk yet. | 12 July 2016, 01:16:38 UTC |
| fb4b1fc | Dave Eargle | 11 July 2016, 17:02:07 UTC | Update retrieving.rst add `mode` to toplevel datastring | 11 July 2016, 17:02:07 UTC |
| d16ecbe | Dave Eargle | 11 July 2016, 16:27:36 UTC | record mode and full useragent string (#240) this stashes the mode in the db (live, sandbox, debug). this can be helpful when doing analyses later down the line in case one needs to distinguish between live and sandbox assignments in the db. the mode is stored as part of the taskdata datastring also, this fixes a neighboring bug where the full useragent string wasn't actually being saved to the taskdata datastring by backbone | 11 July 2016, 16:27:36 UTC |
| cb4eac3 | Dave Eargle | 11 July 2016, 16:26:59 UTC | Fixes #236 (#239) | 11 July 2016, 16:26:59 UTC |
| 1011093 | Dave Eargle | 11 July 2016, 16:26:02 UTC | add_bonus function searches db using uniqueid (#238) Closes #237 | 11 July 2016, 16:26:02 UTC |
| bbd9fc8 | Dave Eargle | 11 July 2016, 16:24:40 UTC | changing the error message for RDS connections for IAM users (#232) * changing the error message for RDS connections for IAM users December 2015 AWS enabled IAM access control for Mturk. This commit edits the error message to be DB-services specific. Right now it makes users think that their credentials aren't good for anything, but actually the message here is only triggered if RDB services can't be accessed. Other commits should maybe change other login messages too to be more specific, but this is a start. Sorry for the obscene length but at least the message is informative! see [this group thread](https://groups.google.com/forum/#!topic/psiturk/1nx8D-tsE74) * simplied the error message, and shunted some of the documentation load for this to the docs. Also linked to a description of using IAM with psiturk. * updating description for openshift and for the mysql description | 11 July 2016, 16:24:40 UTC |
| 6467abe | Dave Eargle | 11 July 2016, 14:39:56 UTC | task parameters doc update (#243) * Update task_parameters.rst with description of experiment_code_version and num_conds bug workaround. | 11 July 2016, 14:39:56 UTC |
| e9c907f | Dave Eargle | 30 June 2016, 00:32:38 UTC | sync url update in comment | 30 June 2016, 00:32:38 UTC |
| ab04063 | Dave Eargle | 27 June 2016, 19:13:38 UTC | Merge pull request #233 from deargle/openshift-mysql-db use openshift db cartridge, if one is installed | 27 June 2016, 19:13:38 UTC |
| 1f9d16e | Dave Eargle | 27 June 2016, 14:07:56 UTC | evading pesky adblockers on ad.html the commonly-used ad-blocking list on https://easylist.github.io/easylist/easylist.txt was recently updated to include 'container-ad', which blocks our entire psiturk ad container. changing our container id. see https://groups.google.com/forum/#!topic/psiturk/yzT-w7KZxms | 27 June 2016, 14:07:56 UTC |
| cf89f8f | Doug Markant | 25 June 2016, 09:20:02 UTC | typo | 25 June 2016, 09:20:02 UTC |
| e362c04 | Dave Eargle | 08 June 2016, 17:02:45 UTC | use openshift db cartridge, if one is installed openshift will host mysql or postgresql relational DBs for you, if you add those cartridges. this commit tells db.py to use the first installed openshift db it finds. note that it is theoretically possible to have both postgresql and mysql addon cartridges installed at the same time. don't do that, that would be silly. No editing config.txt required -- the location of the openshift db can change, so relying on the env vars is really the best way to go. and it's not easy to integrate the env vars into config.txt. See README at https://github.com/deargle/openshift-psiturk-cartridge for an example of getting mysql up and running. | 08 June 2016, 17:02:45 UTC |
| 693e40b | Todd Gureckis | 10 May 2016, 01:17:25 UTC | Merge pull request #231 from deargle/ctrl-c-kill-server notify user when ctrl-c kills experiment server | 10 May 2016, 01:17:25 UTC |
| 796c587 | Todd Gureckis | 10 May 2016, 01:16:50 UTC | Merge pull request #230 from deargle/ctrl-c-new-line Ctrl+c creates a new line | 10 May 2016, 01:16:50 UTC |
| e393d33 | Todd Gureckis | 10 May 2016, 01:16:07 UTC | Merge pull request #209 from deargle/208-get-all-hits get_workers get all hits | 10 May 2016, 01:16:07 UTC |
| e7063ff | David Eargle | 02 May 2016, 19:12:41 UTC | shameful self promotion | 02 May 2016, 19:12:41 UTC |
| 69a711d | David Eargle | 02 May 2016, 18:49:59 UTC | notify user when ctrl-c kills experiment server gunicorn's `on_exit` is hooked so that it can be called when the server is shut down via CTRL+C. Otherwise there is no notification to the user that the server has shut down until they hit `enter` and see that the cmdloop prompt suddenly says "server off" unfortunatley detection of ctrl+c can't be handled reliably by psiturk_shell even by checking `server.is_server_running()` because gunicorn shuts down its workers asynchronously and therefore status will sometimes be read before, during, or after everything has shutdown requires gunicorn >= 19.0 though (when the on_exit hook was implemented). Everything works with the most recent gunicorn (19.4.5) Closes #229 | 02 May 2016, 19:05:12 UTC |
| 704046f | David Eargle | 02 May 2016, 18:44:42 UTC | this function doesn't actually create a blank line -- it is called by _after_ a blank line is submitted. | 02 May 2016, 18:44:42 UTC |
| df80755 | David Eargle | 02 May 2016, 18:31:05 UTC | remove call to parent stub functions each of the removed calls were just stub hook functions in the parend module. no need to call them. | 02 May 2016, 18:36:15 UTC |
| fa752a5 | David Eargle | 02 May 2016, 18:24:24 UTC | create newline on keyboard interrupt. `cmd2`'s `_cmdloop` only exits when `stop` internally evaluates to True. unless there has been a keyboardinterrupt. So if stop is False yet we got here then a keyboard interrupt happened. unfortunately we can't simply listen for a KeyboardInterrupt here because cmd2's `_cmdloop` eats exceptions with a `return` in a `finally` block. so we have to fall back on checking this stop value. we also have to use `_cmdloop()` instead of `cmdloop()` because `cmd2`. | 02 May 2016, 18:24:24 UTC |
| 51158ca | Todd Gureckis | 02 May 2016, 14:58:49 UTC | Merge pull request #228 from deargle/script-no-kill-server this `atkill` handler does _not_ do what it claims. It claims to hand… | 02 May 2016, 14:58:49 UTC |
| cffdae0 | David Eargle | 30 April 2016, 05:15:12 UTC | this `atkill` handler does _not_ do what it claims. It claims to handle keyboard interrupts but that is precisely what the module is designed to not do. what this codeblock _does_ do is kill the server whenever `psiturk -s` is invoked. That's bad. `psiturk -s` should shut down its own servers if that is what the user wants to do. other code already handles shutting down the server on invocation of `exit` or `server off`. this code simply has to go. | 30 April 2016, 05:16:17 UTC |
| a3b9a8f | Todd Gureckis | 26 April 2016, 17:02:37 UTC | Merge pull request #220 from adamliter/fix/219 Adds browser_exclude_rule for Safari | 26 April 2016, 17:02:37 UTC |
| e3806cc | Todd Gureckis | 26 April 2016, 16:26:45 UTC | Merge pull request #218 from NYUCCL/db-timeout-fix fixes "MySQL Server has gone away" error for long-term experiments | 26 April 2016, 16:26:45 UTC |
| 61961fc | Todd Gureckis | 26 April 2016, 16:19:46 UTC | Merge pull request #224 from deargle/patch-2 Describe fractional duration functionality | 26 April 2016, 16:19:46 UTC |
| 4db23e0 | Todd Gureckis | 26 April 2016, 16:19:22 UTC | Merge pull request #225 from deargle/patch-3 build docs in readthedocs theme | 26 April 2016, 16:19:22 UTC |
| f61e477 | Adam Liter | 11 April 2016, 19:12:27 UTC | Adds browser_exclude_rule for Safari This fixes #219. As disscussed in #219, Google Chrome userAgent strings include the word Safari. Because of this, trying to exclude Safari also excludes Google Chrome. The value for Safari userAgent strings just includes Safari and not Google Chrome. Therefore, this problem can be fixed by first checking if both "Chrome" and "Safari" are in the userAgent string. If this is true, then we do not want to exclude it, so we do nothing. If this check returns False, then we still want to check to see if "Safari" is in the userAgent string because that is what we want to exclude. If it is, then we exclude it. | 22 April 2016, 20:52:02 UTC |
| a6777f9 | Dave Eargle | 21 April 2016, 23:36:34 UTC | build docs in readthedocs theme This makes the local build look just like things will look like | 21 April 2016, 23:36:34 UTC |
| 3ae6462 | Dave Eargle | 21 April 2016, 03:29:16 UTC | Describe fractional duration functionality Reflects fractional duration functionality introduced in #180. See also #137, #165. | 21 April 2016, 03:29:16 UTC |
| 418f098 | Todd Gureckis | 18 April 2016, 18:34:48 UTC | Merge pull request #221 from deargle/patch-1 Fixing the readthedocs link to be an external link | 18 April 2016, 18:34:48 UTC |
| e31484e | Todd Gureckis | 18 April 2016, 17:50:15 UTC | to address #222: comment out imgmath extension from sphynx (build errors on readthedocs) | 18 April 2016, 17:50:15 UTC |
| 1a787b6 | Todd Gureckis | 18 April 2016, 17:42:23 UTC | Merge pull request #223 from deargle/docs-external-survey Docs for how to integrate external survey tools such as Qualtrics | 18 April 2016, 17:42:23 UTC |
| 040409c | Dave Eargle | 17 April 2016, 05:41:23 UTC | the new page and its fine linkages | 17 April 2016, 06:08:10 UTC |
