-
Notifications
You must be signed in to change notification settings - Fork 25
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[WIP] dandi search command #1126
base: master
Are you sure you want to change the base?
Changes from all commits
c2fb996
bd210e0
2852e42
447c805
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,212 @@ | ||
from pathlib import Path | ||
|
||
from SPARQLWrapper import JSON, SPARQLWrapper | ||
import click | ||
import pandas as pd | ||
|
||
from .base import map_to_click_exceptions | ||
|
||
# supported fields | ||
# TODO: adding more | ||
DANDISETS_FIELDS = { | ||
"approach": ["apr", "?as dandi:approach / schema:name ?apr ."], | ||
"species_id": ["sid", "?as dandi:species / schema:identifier ?sid ."], | ||
"species_name": ["snm", "?as dandi:species / schema:name ?snm ."], | ||
} | ||
|
||
ASSETS_FIELDS = { | ||
"size": ["size", "?asset schema:contentSize ?size ."], | ||
"format": ["format", "?asset schema:encodingFormat ?format ."], | ||
"age": ["age", "?asset prov:wasAttributedTo / dandi:age / schema:value ?age ."], | ||
} | ||
|
||
|
||
def parse_validate(ctx, param, value): | ||
value_parse = [] | ||
# parsing elements that have multiple comma-separated values | ||
for el in value: | ||
value_parse += el.split(",") | ||
if param.name == "select_fields": | ||
if ctx.params["search_type"] == "dandisets": | ||
choice_list = DANDISETS_FIELDS.keys() | ||
elif ctx.params["search_type"] == "assets": | ||
choice_list = ASSETS_FIELDS.keys() | ||
else: | ||
choice_list = None | ||
else: | ||
choice_list = None | ||
# checking if all values are in the list of possible choices | ||
for el in value_parse: | ||
if choice_list and el not in choice_list: | ||
ctx.fail(f"{el} is not in the list: {choice_list}") | ||
return value_parse | ||
|
||
|
||
@click.command(help="Search TODO") | ||
@click.option( | ||
"-F", | ||
"--file", | ||
help="Comma-separated list of fields to display. " | ||
"An empty value to trigger a list of " | ||
"available fields to be printed out", | ||
) | ||
@click.option( | ||
"-t", | ||
"--search_type", | ||
help="Type of the search.", | ||
type=click.Choice(["dandisets", "assets"]), | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. familiarize with #1357 and RF similarly here |
||
) | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. FTR we have
I wonder if we should unify somehow |
||
@click.option( | ||
"-s", | ||
"--select_fields", | ||
help="Field name for dandisets search", | ||
callback=parse_validate, | ||
multiple=True, | ||
) | ||
@click.option( | ||
"-f", | ||
"--filter_fields", | ||
help="Field name for dandisets search", | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. help is the same as for There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. also we have in ls
I think here we should be uniform with that -- with dandi/dandi-archive#1659 in mind, we do adhere to MVC pattern here too, and clearly separate out the model from rendering and thus have similar |
||
type=(str, str), | ||
multiple=True, | ||
) | ||
@click.option( | ||
"--format", | ||
help="Choose the format for output. TODO", | ||
type=click.Choice(["stdout", "csv"]), | ||
default="stdout", | ||
) | ||
@click.option( | ||
"--number_of_lines", | ||
help="Number of lines of output that will be printed", | ||
default=10, | ||
) | ||
@click.option( | ||
"-d", | ||
"--database_name", | ||
help="Database name", | ||
default="dandisets_new", | ||
) | ||
@map_to_click_exceptions | ||
def search( | ||
file=None, | ||
search_type=None, | ||
select_fields=None, | ||
filter_fields=None, | ||
format="stdout", | ||
number_of_lines=10, | ||
database_name="dandisets_new", | ||
): | ||
|
||
if file and search_type: | ||
raise Exception("file and type are mutually exclusive options") | ||
|
||
if file: | ||
filepath = Path(file) | ||
with filepath.open() as f: | ||
query_str = f.read() | ||
elif search_type in ["dandisets", "assets"]: | ||
if not select_fields: | ||
raise Exception( | ||
f"select_fields is required if search type is {search_type}" | ||
) | ||
if filter_fields: | ||
for el in filter_fields: | ||
if el[0] not in select_fields: | ||
raise Exception( | ||
f"field {el[0]} used in filter_fields, " | ||
f"but select fields contain {select_fields}" | ||
) | ||
if search_type == "dandisets": | ||
query_str = create_dandisets_query(select_fields, filter_fields) | ||
elif search_type == "assets": | ||
query_str = create_assets_query(select_fields, filter_fields) | ||
else: | ||
raise NotImplementedError | ||
|
||
endpoint = "https://search.dandiarchive.org:5820/dandisets_new/query" | ||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. this would be just 1 type of search. We already have a few more (even if not exposed via API): https://llmsearch.dandiarchive.org/, https://dandiarchive.org/search etc. I think we should generalize this |
||
sparql = SPARQLWrapper(endpoint) | ||
sparql.setCredentials("anonymous", "anonymous") | ||
sparql.setReturnFormat(JSON) | ||
sparql.setQuery(query_str) | ||
results = sparql.queryAndConvert() | ||
res_df = results2df(results, number_of_lines) | ||
|
||
if format == "stdout": | ||
print(res_df) | ||
else: | ||
raise NotImplementedError("only stdout format implemented for now") | ||
|
||
|
||
def results2df(results, limit=10): | ||
res_lim = results["results"]["bindings"][:limit] | ||
res_val_l = [dict((k, v["value"]) for k, v in res.items()) for res in res_lim] | ||
return pd.DataFrame(res_val_l) | ||
|
||
|
||
def filter_query(filter_fields, fields_dict): | ||
"""creating filter part for the queries""" | ||
filter_str = "" | ||
for (key, val) in filter_fields: | ||
if val[0] == "(" and val[-1] == ")": | ||
val = val[1:-1].split(",") | ||
if len(val) != 2: | ||
raise ValueError( | ||
"If value for filter is a tuple, it has to have 2 elements " | ||
) | ||
else: | ||
min_val = val[0].strip() | ||
max_val = val[1].strip() | ||
if max_val and min_val: | ||
filter_str += ( | ||
f"FILTER (?{fields_dict[key][0]} > {min_val} " | ||
f"&& ?{fields_dict[key][0]} < {max_val}) \n" | ||
) | ||
elif max_val: | ||
filter_str += f"FILTER (?{fields_dict[key][0]} < {max_val}) \n" | ||
elif min_val: | ||
filter_str += f"FILTER (?{fields_dict[key][0]} > {min_val}) \n" | ||
else: | ||
val = val.split(",") | ||
cond_str = f'?{fields_dict[key][0]} = "{val[0]}"' | ||
for el in val[1:]: | ||
cond_str += f' || ?{fields_dict[key][0]} = "{el}"' | ||
filter_str += f"FILTER ({cond_str}) \n" | ||
return filter_str | ||
|
||
|
||
def create_dandisets_query(select_fields, filter_fields): | ||
"""Creating a query for dandisets search""" | ||
var = "" | ||
for el in select_fields: | ||
var += f" ?{DANDISETS_FIELDS[el][0]}" | ||
|
||
query_str = ( | ||
f"SELECT DISTINCT ?d{var} WHERE \n" "{ \n" " ?d dandi:assetsSummary ?as . \n" | ||
) | ||
for el in select_fields: | ||
query_str += f" {DANDISETS_FIELDS[el][1]} \n" | ||
query_str += filter_query(filter_fields, DANDISETS_FIELDS) | ||
query_str += "}" | ||
return query_str | ||
|
||
|
||
def create_assets_query(select_fields, filter_fields): | ||
"""Creating a query for assets search""" | ||
var = "" | ||
for el in select_fields: | ||
var += f" ?{ASSETS_FIELDS[el][0]}" | ||
|
||
query_str = ( | ||
f"SELECT DISTINCT ?asset ?d_id ?path{var} WHERE \n" | ||
"{ \n" | ||
" ?asset rdf:type dandi:Asset . \n" | ||
" ?d prov:hasMember ?asset . \n" | ||
" ?d schema:identifier ?d_id . \n" | ||
" ?asset dandi:path ?path . \n" | ||
) | ||
for el in select_fields: | ||
query_str += f" {ASSETS_FIELDS[el][1]} \n" | ||
query_str += filter_query(filter_fields, ASSETS_FIELDS) | ||
query_str += "}" | ||
return query_str |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
here and everywhere below use
--long-option
, not--long_option
as we do everywhere else (rungit grep -A5 '@click.option'
)