Location to/from locator mapping facilities #661
Conversation
Change the `MathJax.Hub.Queue(initializeRequest)` to a simpler function call (`initializeRequest.call(this)`). This was failing to give a proper context to initializeRequest, and `this.value` was turning up as `undefined`. Also add a fallback if we need to display some code before MathJax finishes its original typesetting. I was stubbing out `Queue` in my specs, so the tests had to be changed around a little.
… on the base method that is in xml_module.py to preserve the correct export filesystem hierarchy.
…ch presumes writing to a filesystem, plus it adds a url_name to the attribute set. Also, on __init__ reset the 'category' attribute, on some code paths this can get lost. Not sure why, but this gets all the tests to pass.
Also fixes problems with double-quoted strings (historical artifact). STUD-640 Conflicts: common/lib/xmodule/xmodule/video_module.py
…e XModuleDescriptor
The django_nose it is very useful, even outside the test environment. For example, it lets you to easily run test from manage.py without additional changes to the test packages.
Fix grade range names not being editable.
Video: disable speed controls for unsupported browsers.
| a singleton accessor. | ||
| """ | ||
| # pylint: disable=W0603 | ||
| global _loc_singleton |
There was a problem hiding this comment.
Rather than ignoring the pylint errors, why not just make this _LOC_SINGLETON (or, even better, LOCATION_MAPPER)
There was a problem hiding this comment.
I didn't know those would suffice for pylint. I'm wary of the
non-underscored version as implying "public" accessibility.
On Fri, Aug 16, 2013 at 9:36 AM, Calen Pennington
notifications@github.comwrote:
In common/lib/xmodule/xmodule/modulestore/django.py:
@@ -46,3 +47,18 @@ def modulestore(name='default'):
# Initialize the modulestores immediately
for store_name in settings.MODULESTORE:
modulestore(store_name)
+_loc_singleton = None
+def loc_mapper():
- """
- Get the loc mapper which bidirectionally maps Locations to Locators. Used like modulestore() as
- a singleton accessor.
- """
pylint: disable=W0603
- global _loc_singleton
Rather than ignoring the pylint errors, why not just make this
_LOC_SINGLETON (or, even better, LOCATION_MAPPER)—
Reply to this email directly or view it on GitHubhttps://github.com/edx/edx-platform/pull/661/files#r5813741
.
There was a problem hiding this comment.
Changing case didn't get rid of the warning. The difference between this and _MODULESTORE above is that this does a direct assignment v changing something inside of it which triggers the need for the global decl, afaik.
There was a problem hiding this comment.
I'm ok w/ disabling the complaint about global. I was more concerned about a complaint about casing.
There was a problem hiding this comment.
W0603 http://pylint-messages.wikidot.com/messages:w0603: Using the global
statement (from http://pylint-messages.wikidot.com/all-messages)
On Fri, Aug 16, 2013 at 10:59 AM, Calen Pennington <notifications@github.com
wrote:
In common/lib/xmodule/xmodule/modulestore/django.py:
@@ -46,3 +47,18 @@ def modulestore(name='default'):
# Initialize the modulestores immediately
for store_name in settings.MODULESTORE:
modulestore(store_name)
+_loc_singleton = None
+def loc_mapper():
- """
- Get the loc mapper which bidirectionally maps Locations to Locators. Used like modulestore() as
- a singleton accessor.
- """
pylint: disable=W0603
- global _loc_singleton
I'm ok w/ disabling the complaint about global. I was more concerned about
a complaint about casing.—
Reply to this email directly or view it on GitHubhttps://github.com/edx/edx-platform/pull/661/files#r5815544
.
…yback. Because with speed changing the time also changes for Flash playback, a different way to calculate the current time is needed than for HTML5 playback. I have added conditions for Flash and HTML5 video, and put old method of calculating time for Flash. I have tested it on the YouTube video ZwkTiUPN0mg. Both HTML5 mode and Flash mode have proper video-captions syncing with this fix. NOTE: to view YouTube video in Flash mode you either have to use an old browser (ex. Firefox version 18) or hard code in source that state.currentPlayerMode = 'flash' (in function _setPlayerMode(), file 01_initialize.js).
| self.create_map_entry(course_location) | ||
| entry = self.location_map.find_one(location_id) | ||
| else: | ||
| return None |
There was a problem hiding this comment.
This seems like a reasonable place to raise an exception, rather than just returning None
|
At a larger scale, it seems like this mapping facility might be better suited to storage in a relational database. Did you consider doing it that way? |
in grading, if problem cannot be created, return score as none
Letting xblocks handle scope rather than separating fields into different attrs. Although, split still shunts content fields to a different collection than setting and children fields. The big difference is that content fields will always be a dict and not sometimes just a string and there's no special casing of 'data' attr. The other mind change is no more 'metadata' dict.
and fix some comments ref'g old names
Correctly seek in captions with non-1.0 speeds.
|
I didn't consider sql b/c I'd assumed that someone made the architectural decision that all course content info belongs in mongo and all user id and progress belongs in sql (transactional data). afaik, there's no objective reason for preferring mongo over sql for any of our data since we don't use any of the doc processing capabilities of mongo (CLOBs could store our definitions just as well in a sql db.) |
xblock fields persist w/o breaking by scope
Add a 'MixedModuleStore' which aggregates between XML and Mongo Modulestores
|
Addressed everything but the n^2 name generator (which I addressed w/ sophistry). |
per our agreement not to use the domain
partial impl
unit tests and fixes they implied
But don't use this parser for urls embedded in a browser url request.
with own db connection.
…estore() singleton accessor
and some minor comment changes.
|
rebased |
|
👍 |
|
The reason we put course content in mongo was so that we could search inside the content at some point in the future (which we couldn't do w/ CLOBs). That doesn't mean that all things course related have to go in mongo. In general, code should use a data store that matches its need. In this case, I'm not sure what it would take to put this in a sql store (for instance, would it introduce django dependencies into the modulestore codebase, or would we move this all out of modulestore into common/djangoapps, or would we add a separate sql mapper...) In the end, I think let's leave it in mongo for now, and if we have trouble w/ performance, we can migrate it to sql later. |
Location to/from locator mapping facilities
|
@dmitchell @cpennington this PR made master unstable with too many pep8/pylint violations. |
Fix circuit simulator js for Firefox
…-merge-achievement Implements ga_achievement openedx#578, openedx#594
Hash update: problem builder.
feat: backport filters
was pr 601 but I accidentally blew it away when I rebased --onto master w/ push -f afterward
@cpennington