Add Search Studies object #190
Conversation
|
|
||
| Examples | ||
| -------- | ||
| Searches are done using natural language, with AND, OR, and NOT supported, as |
There was a problem hiding this comment.
I don't think this qualifies as natural language (i.e., "plain english")
There was a problem hiding this comment.
Yeah, my closest guess would be a query language then.
There was a problem hiding this comment.
Boolean?
On Jul 9, 2014 7:28 PM, "Joshua Shorenstein" notifications@github.com
wrote:
In qiita_db/search.py:
+This module provides functionality for searching studies and samples contained
+in the qiita database. All language processing and querying of the database is
+contained within each object.
+
+Classes
+-------
+
+..autosummary::
- :toctree: generated/
- QiitaStudySearch
+Examples
+--------
+Searches are done using natural language, with AND, OR, and NOT supported, asYeah, my closest guess would be a query language then.
—
Reply to this email directly or view it on GitHub
https://github.com/biocore/qiita/pull/190/files#r14746084.
There was a problem hiding this comment.
Yup, that looks like it.
|
Are we planning to also add (at some point) an |
|
I'd like to see it, probably add an issue so we can keep track of it? |
| # create the sample finding SQL | ||
| sample_sql = ("SELECT r.sample_id FROM qiita.required_sample_info r " | ||
| "JOIN qiita.sample_%s s ON s.sample_id = r.sample_id " | ||
| "WHERE {0}".format(sql_where)) |
There was a problem hiding this comment.
If the above INTERSECTs do not find any tables, this will throw an SQL error, right?
There was a problem hiding this comment.
Possible to join on INTERSECT instead of magic numbers?
On Jul 9, 2014 5:44 PM, "adamrp" notifications@github.com wrote:
In qiita_db/search.py:
# create the study finding SQL# remove metadata headers that are in required_sample_info tablemeta_headers = meta_headers.difference(required_cols)# get all study ids that contain all metadata categories searched forsql = []for meta in meta_headers:sql.append("SELECT study_id FROM qiita.study_sample_columns WHERE ""column_name = '%s' INTERSECT" %scrub_data(meta))# combine the query, stripping off the last INTERSECTstudy_sql = ' '.join(sql)[:-10]
# create the sample finding SQLsample_sql = ("SELECT r.sample_id FROM qiita.required_sample_info r ""JOIN qiita.sample_%s s ON s.sample_id = r.sample_id ""WHERE {0}".format(sql_where))If the above INTERSECTs do not find any tables, this will throw an SQL
error, right?—
Reply to this email directly or view it on GitHub
https://github.com/biocore/qiita/pull/190/files#r14743256.
|
This looks pretty awesome! I gave it a somewhat quick review during a talk, so it would be good to get more eyes on it. |
|
@josenavas, created #191 |
|
Thanks! |
|
After starting work on the web front-end for this I realized some features were missing. I've added them in this pull request, so it needs another thorough looking over. |
|
I reviewed this with @squirrelo and there were just a couple things 1) an idea on how to get the tests to work with the global variable, and 2) add a test for a complex query with nested/composed ORs and ANDs to make sure that everything is being generated properly. After that, I give it a 👍 and he just needs someone else to go over it for a merge! Thanks @squirrelo |
| results = {} | ||
| # run search on each study to get out the matching samples | ||
| for sid in study_ids: | ||
|
|
There was a problem hiding this comment.
Does this blank line need to be here?
This object has natural language parsing for searches and returns a dictionary of {study_id: [samp1, samp2, ...]} of samples and studies matching the search query. An example query is also in the documentation, showing you can get pretty complex with them.