Skip to content

Commit

Permalink
Merge pull request #1157 from girder/better-column-filter
Browse files Browse the repository at this point in the history 
Allow more sophisticated filtering on item lists.
  • Loading branch information
@manthey
manthey authored May 15, 2023
2 parents d8f9bc7 + f9bd8dd commit 1e6c103
Show file tree
Hide file tree
Showing 2 changed files with 71 additions and 31 deletions.
2 changes: 2 additions & 0 deletions girder/girder_large_image/rest/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,8 @@ def altItemFind(self, folderId, text, name, limit, offset, sort, filters=None):
text = None
except Exception as exc:
logger.warning('Failed to parse _filter_ from text field: %r', exc)
if filters:
logger.debug('Item find filters: %s', json.dumps(filters))
if recurse:
return _itemFindRecursive(
self, origItemFind, folderId, text, name, limit, offset, sort, filters)
Expand Down
100 changes: 69 additions & 31 deletions girder/girder_large_image/web_client/views/itemList.js
View file
Original file line number Diff line number Diff line change
Expand Up @@ -219,50 +219,88 @@ wrap(ItemListWidget, 'render', function (render) {
addToRoute({filter: this._generalFilter});
};

this._unescapePhrase = (val) => {
if (val !== undefined) {
val = val.replace('\\\'', '\'').replace('\\"', '"').replace('\\\\', '\\');
}
return val;
};

this._setFilter = () => {
const val = this._generalFilter;
let filter;
const usedPhrases = {};
const columns = (this._confList() || {}).columns || [];
if (val !== undefined && val !== '' && columns.length) {
// a value can be surrounded by single or double quotes, which will
// be removed.
const quotedValue = /((?:"((?:[^\\"]|\\\\|\\")*)"|'((?:[^\\']|\\\\|\\')*)'|([^:,\s]+)))/g;
const phraseRE = new RegExp(
new RegExp('((?:' + quotedValue.source + ':|))').source +
/(-?)/.source +
quotedValue.source +
new RegExp('((?:,' + quotedValue.source + ')*)').source, 'g');
filter = [];
val.match(/"[^"]*"|'[^']*'|\S+/g).forEach((phrase) => {
if (!phrase.length || usedPhrases[phrase]) {
return;
[...val.matchAll(phraseRE)].forEach((match) => {
const coltag = this._unescapePhrase(match[5] || match[4] || match[3]);
const phrase = this._unescapePhrase(match[10] || match[9] || match[8]);
const negation = match[6] === '-';
var phrases = [phrase];
if (match[11]) {
[...match[11].matchAll(quotedValue)].forEach((submatch) => {
const subphrase = this._unescapePhrase(submatch[4] || submatch[3] || submatch[2]);
if (subphrase && subphrase.length && !phrases.includes(subphrase)) {
phrases.push(subphrase);
}
});
}
usedPhrases[phrase] = true;
if (phrase[0] === phrase.substr(phrase.length - 1) && ['"', "'"].includes(phrase[0])) {
phrase = phrase.substr(1, phrase.length - 2);
const key = `${coltag}:` + phrases.join('|||');
if (!phrases.length || usedPhrases[key]) {
return;
}
const numval = +phrase;
/* If numval is a non-zero number not in exponential notation.
* delta is the value of one for the least significant digit.
* This will be NaN if phrase is not a number. */
const delta = Math.abs(+numval.toString().replace(/\d(?=.*[1-9](0*\.|)0*$)/g, '0').replace(/[1-9]/, '1'));
// escape for regex
phrase = phrase.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
usedPhrases[key] = true;
const clause = [];
columns.forEach((col) => {
let key;
phrases.forEach((phrase) => {
const numval = +phrase;
/* If numval is a non-zero number not in exponential
* notation, delta is the value of one for the least
* significant digit. This will be NaN if phrase is not a
* number. */
const delta = Math.abs(+numval.toString().replace(/\d(?=.*[1-9](0*\.|)0*$)/g, '0').replace(/[1-9]/, '1'));
// escape for regex
phrase = phrase.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');

if (col.type === 'record' && col.value !== 'controls') {
key = col.value;
} else if (col.type === 'metadata') {
key = 'meta.' + col.value;
}
if (key) {
clause.push({[key]: {$regex: phrase, $options: 'i'}});
if (!_.isNaN(numval)) {
clause.push({[key]: {$eq: numval}});
if (numval > 0 && delta) {
clause.push({[key]: {$gte: numval, $lt: numval + delta}});
} else if (numval < 0 && delta) {
clause.push({[key]: {$lte: numval, $gt: numval + delta}});
columns.forEach((col) => {
let key;

if (coltag && coltag !== col.value) {
// do we want to match the last .<fragment> as well?a
// do we want to be case insensitive?
return;
}
if (col.type === 'record' && col.value !== 'controls') {
key = col.value;
} else if (col.type === 'metadata') {
key = 'meta.' + col.value;
}
if (key) {
clause.push({[key]: {$regex: phrase, $options: 'i'}});
if (!_.isNaN(numval)) {
clause.push({[key]: {$eq: numval}});
if (numval > 0 && delta) {
clause.push({[key]: {$gte: numval, $lt: numval + delta}});
} else if (numval < 0 && delta) {
clause.push({[key]: {$lte: numval, $gt: numval + delta}});
}
}
}
}
});
});
filter.push({$or: clause});
if (clause.length > 0) {
filter.push(!negation ? {$or: clause} : {$nor: clause});
} else if (!negation) {
filter.push({$or: [{_no_such_value_: '_no_such_value_'}]});
}
});
if (filter.length === 0) {
filter = undefined;
Expand Down Expand Up @@ -296,7 +334,7 @@ wrap(ItemListWidget, 'render', function (render) {
func = 'before';
}
if (base.length) {
base[func]('<span class="li-item-list-filter">Filter: <input class="li-item-list-filter-input""></input></span>');
base[func]('<span class="li-item-list-filter">Filter: <input class="li-item-list-filter-input" title="All specified terms must be included. Surround with quotes to include spaces. Prefix with - to exclude that value. By default, all columns are searched. Use <column>:<value1>[,<value2>...] to require that a column match a specified value or any of a list of specified values. Column and value names can be quotes to include spaces. If <column>:-<value1>,[.<value2>] is specified, matches will exclude the list of values."></input></span>');
if (this._generalFilter) {
root.find('.li-item-list-filter-input').val(this._generalFilter);
}
Expand Down

0 comments on commit 1e6c103

Please sign in to comment.