feat: expose functionality to extract unit content #37
Conversation
| return None | ||
|
|
||
|
|
||
| def _leaf_filter(block): |
There was a problem hiding this comment.
This is a custom leaf filter to exclude blocks that are not video or html, that way we limit any calculations/processing on nodes that aren't relevant to what we need.
| length - the cummulative length of a block's children's content | ||
| items - a list of dictionaries containing the content type and text for each child | ||
| """ | ||
| block = get_single_block(request, user_id, course_id, unit_usage_key) |
There was a problem hiding this comment.
I'm only using load_single_xblock once here on the parent block. I wasn't sure if I need to do that on each child block to check visibility.
There was a problem hiding this comment.
You shouldn't need to, no. load_single_xblock should only return the children that the learner has access to, so long as will_recheck_access=False.
alangsto
force-pushed
the
alangsto/extract_unit_content
branch
from
4d5653e to
b2b8e6e
Compare
| return length, items | ||
|
|
||
|
|
||
| def get_block_content(request, user_id, course_id, unit_usage_key): |
There was a problem hiding this comment.
for my own understanding, what calls this?
There was a problem hiding this comment.
This function will be used as part of https://2u-internal.atlassian.net/browse/COSMO-73 - essentially whatever code is putting together the prompt will call it.
| return transcript | ||
|
|
||
|
|
||
| def get_single_block(request, user_id, course_id, usage_key_string, course=None, will_recheck_access=False): |
There was a problem hiding this comment.
I don't think we'll ever want this to be True because then we won't be checking learner access to the block and its children. I think it would be better not to expose this function parameter in that case, since we don't want someone to accidentally pass True. What do you think?
There was a problem hiding this comment.
Oh, that's a great point! I'll update this function.
| length - the cummulative length of a block's children's content | ||
| items - a list of dictionaries containing the content type and text for each child | ||
| """ | ||
| block = get_single_block(request, user_id, course_id, unit_usage_key) |
There was a problem hiding this comment.
You shouldn't need to, no. load_single_xblock should only return the children that the learner has access to, so long as will_recheck_access=False.
| """ | ||
| Given a specific block, return the content type and text of a pre-order traversal of the blocks children. | ||
| """ | ||
| leaf_nodes = traverse_block_pre_order(block, block_get_children, _leaf_filter) |
| self.course_id = 'course-v1:edx+test+23' | ||
|
|
||
| @ddt.data( | ||
| ('video', True), |
There was a problem hiding this comment.
Should we include html here for completeness?
There was a problem hiding this comment.
Yes, thanks for catching that!
alangsto
force-pushed
the
alangsto/extract_unit_content
branch
from
b2b8e6e to
69f8945
Compare
alangsto
force-pushed
the
alangsto/extract_unit_content
branch
from
69f8945 to
285e0e2
Compare
COSMO-70
In order to provide unit specific content to a prompt template, the content needs to be extracted from a given block. Pieces of this code have been taken from the ai_aside repository, specifically the
text_utils.pyfile.Content from a block is retrieved using a pre-order traversal, meaning that all content will be returned in the order a learner would be viewing the content.