Filter blocks UI (1) #247

alexkb0009 · 2020-10-22T19:13:02Z

(Outdated Pic, more features/etc now:)

Current Implementation Notes

Decided to re-implement the 'ActiveFiltersBar' since might be able to create better / more-responsive layouting using flexboxes now. (And is not really much code, esp JS-side). (Not fully complete).

Next Challenges

Will need to work on FilterSetController to:

~~store currently selected FilterSet Block~~ - done mostly
~~if multiple selected then don't show FacetList~~
- ~~(De)Selection will occur on click (?)~~
- Update: for now for simplicity, allowing only 1 (or 0 (aka 'all')) blocks to be selected.
~~if 1 selected, then show FacetList AND display current search query/filters rather than saved one (since we'll save this, if we're editing)~~
~~Will likely hold a 'not yet saved' FilterSet Item in memory and apply changes to it as they're made in FacetList... until saved..~~
- Done & done
~~Need to figure out intricacies of above~~

All Completed.

Remaining Fixes

Need to (a) fix CustomColumnSelector/Controller re: compound filter sets (and EmbeddedSearchViews as a whole) and Done in SPC master, will update SPC version to 0.0.2.83 before merging. (b) ~~figure out if we save 'sort' param to filterset or similar (probably not.. for time being at least)~~ sort is currently not saved anywhere on FilterSet, we completely lose 'sort' if change to a different filterblock/compound-request.

Possible ToDos

~~AJAX in active_filterset by its @id to avoid needing to embed extra fields on Case...~~ completed.
- This might be simpler to do after rest of stuff is complete, including searchbox for selecting previous/other filtersets; could perhaps optimize to re-use search result if we request limit=all or sort=date_created&sort=actively_used something...

… filter-blocks-ui-1

TODO: Consider AJAXing it initially? Hmm..

… filter-blocks-ui-1

willronchetti

Some preliminary comments/questions that I hope are helpful as you continue to work on this.

willronchetti · 2020-10-30T18:42:44Z

src/encoded/search/compound_search.py

+        # sub_request.environ = env  # copy back in
        sub_request.__parent__ = None  # XXX: set parent request (is None *always* correct?) -Will
        sub_request.registry = parent_request.registry  # transfer registry as well
+        if hasattr(parent_request, "context"):
+            sub_request.context = parent_request.context
+        else:
+            sub_request.context = None


Let's go ahead and persist this change - you can get rid of the commented out code.

willronchetti · 2020-10-30T18:43:12Z

src/encoded/search/compound_search.py

+
+        #env = parent_request.environ.copy()  # copy parent request
+        #env['PATH_INFO'] = sub_request.environ['PATH_INFO']  # override what it does
+        #env['QUERY_STRING'] = sub_request.environ['QUERY_STRING']
+
        if '?' not in query:  # do some sanitization
            query = '?' + query
-        subreq = Request.blank(route + '%s&from=%s&limit=%s' % (query, from_, to))
+        # subreq = Request.blank(route + '%s&from=%s&limit=%s' % (query, from_, to))
+
+        subreq = make_subrequest(request, route + '%s&from=%s&limit=%s' % (query, from_, to), "GET")
+        subreq.headers['Accept'] = 'application/json'
+


Same deal here - remove commented out code, keep your change

willronchetti · 2020-10-30T18:45:34Z

src/encoded/search/search_utils.py

+
+def build_initial_columns(used_type_schemas):
+
+    columns = OrderedDict()


All Python 3.6 dictionaries are OrderedDicts, so can just use columns = {} but can keep the specificity if you want.

willronchetti · 2020-10-30T18:46:11Z

src/encoded/search/search_utils.py

@@ -371,3 +372,43 @@ def is_numerical_field(field_schema):
 def is_date_field(field, field_schema):
    """ Helper method that determines if field_schema is  """
    return determine_if_is_date_field(field, field_schema)
+
+def build_initial_columns(used_type_schemas):


This I think should be a @staticmethod on SearchBuilder, unless you think this has external use?

Not quite sure but probably not... yeah staticmethod makes sense for it.. will change

willronchetti · 2020-10-30T18:52:41Z

src/encoded/search/search_utils.py

+                    # If @type or display_title etc. column defined in schema, then override defaults.
+                    for prop in schema_columns[name]:
+                        columns[name][prop] = schema_columns[name][prop]


I think columns[name].update(schema_columns[name]) is what you want here?

willronchetti · 2020-11-02T15:53:13Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+    //     // TODO: Show single > or < or something.
+    // }
+
+    const fieldSchema = getSchemaProperty(field, Schemas.get(), "VariantSample");


Should we bake assumptions in like this? It would be ideal if filter_sets were generic client side as well that way they can be used for other things (in the future potentially).

In long run yeah I think that might be good idea but for time being was working under impression FilterSetUI would only apply to table of variant samples so left some things hardcoded re: that since results in less complex code. Once/if we do need to repurpose for other search_types, it shouldn't be much issue to parameterize-upon / allow-dynamic-value-for it though.

willronchetti · 2020-11-02T15:53:31Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+            "@id" : PropTypes.string, // Is required if originally existed, else free to be null.
+            "uuid" : PropTypes.string, // Is required if originally existed, else free to be null.
+            "title" : PropTypes.string.isRequired,
+            "search_type" : PropTypes.oneOf(["VariantSample", "Variant", "Case"]),


Ditto here.

PropTypes don't throw any hard errors, but do generate warnings in browser console in dev environments. Generally we use them as if were documentation/reminders for type definitions in place of like JSDoc.

willronchetti · 2020-11-02T15:55:02Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+        const excludedFieldMap = {};
+        if (Array.isArray(excludeFacets)) {
+            excludeFacets.forEach(function(field){
+                excludedFieldMap[field] = true;
+            });
+        }


Why not just make this a dictionary from the onset?

Good point.. legacy compatibility mostly... this might be a good opportunity to change it, though also hasn't been much of an issue. Will try to maybe do depending on time/other-stuff.

willronchetti · 2020-11-02T15:56:20Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+        this.addNewFilterBlock = _.throttle(this.addNewFilterBlock.bind(this), 750, { trailing: false });
+        // Throttled, but func itself throttles/prevents-update if still loading last-selected search results.
+        this.selectFilterBlockIdx = _.throttle(this.selectFilterBlockIdx.bind(this), 250, { trailing: false });
+        // Throttled to prevent accidental double-clicks.
+        this.removeFilterBlockAtIdx =  _.throttle(this.removeFilterBlockAtIdx.bind(this), 250, { trailing: false });


Same question as previously WRT throttle values.

Mix of 'low enough to not impact usability during functional/real usage' and 'high enough to prevent people rapidly clicking it a lot and driving up server load with lot of search requests'. (Almost all of these methods ultimately trigger search query).

willronchetti · 2020-11-02T15:57:12Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+
+        this.state = {
+            "currFilterSet" : { ...initialFilterSetItem },
+            // todo: maybe change to allow multiple in future?


I definitely think this is a feature that will be requested ie: quickly swap from filter_set_1 to filter_set_2 etc.

… filter-blocks-ui-1

This reverts commit 44d3218.

… filter-blocks-ui-1

Bianca-Morris

Still haven't tested this on my local machine, so don't want to nitpick too much until I understand more of how this works/looks. But will soon (test, not necessarily nitpick, ha). Is there a related branch on SPC I need to test it with?

Bianca-Morris · 2020-11-19T02:39:43Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+        const duplicateNameIndices = {};
+
+        filter_blocks.forEach(function({ name, query }, idx){
+            var i;


Ah, this can be confusing. var can be used to make a variable global, if it's defined in global scope. But here, it's defined inside of a function, so i is local to the function, but isn't available outside of it. let is typically scoped to the block (for loop, conditional statement, etc.), but since it would be defined outside of the loops here, that wouldn't make much of a difference. So, var is fine in this case, but let would work, too.

Typically better to use let to avoid bugs unless there's a scope-related reason to use var. So you're actually kinda right, but it's for a different reason.

Bianca-Morris · 2020-11-19T02:52:04Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+        if ( // Rebuild tooltips after stuff that affects tooltips changes.
+            pastFilterSet !== currFilterSet || // If FilterSet changes, then some tips likely for it do as well, esp re: validation/duplicate-queries.
+            selectedFilterBlockIdx !== pastSelectedIdx ||
+            (bodyOpen && !pastBodyOpen) || // `data-tip` elems not visible until body mounted in DOM
+            cachedCounts !== pastCachedCounts // Tooltip on filterblocks' counts indicator, if present.
+        ) {
+            setTimeout(ReactTooltip.rebuild, 0);
+        }
+    }


Ah, thanks for reminding me how to do this. I keep running into tooltip-related woes, and forgot you could do a rebuild.

A small consideration in back of my head has been what's performance of these ReactTooltip.rebuild calls. By its nature, I think it traverses the entire DOM tree to find these 'data-tip' attributes, which does not seem it might be necessarily fast.

Just profiled it out of curiosity and seems it only takes about 3ms though, during otherwise idle time (thanks to the setTimeout), not super fast but good. One random idea if it were an issue might be to debounce the ReactTooltip.rebuild func so if called multiple times within some period of time, gets called only once at tail end of that period.

Bianca-Morris · 2020-11-19T03:05:35Z

src/encoded/static/components/item-pages/components/EmbeddedItemSearchTable.js

+EmbeddedItemSearchTable.defaultProps = {
+    "columnExtensionMap": columnExtensionMapCGAP,
+    "facets" : undefined // Default to those from search response.
+};


Why explicitly set facets prop to undefined? Won't props typically default to that if nothing is passed through, anyway?

Just for explicitness / documentation-for-self mostly. I figured if is explicitly set to undefined then "undefined" feels more of a valid value, rather than having it feel like some accidental/overlooked lack of value which then feels like some sort of issue to be solved (especially if come back to this code after long time of not working on it) rather than an explicit design choice that allows/handles "undefined"... idk. Lmk if other thoughts, am def up to changing this.. and/or documenting.. etc.

Bianca-Morris · 2020-11-19T03:13:48Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+            if (!React.isValidElement(child)) { // String or something
+                return child;
+            }
+            if (typeof child.type === "string") { // Normal element (a, div, etc)
+                return child;
+            }


Why not merge these two into one if ( x || y ) statement? If you wanted to keep the comments, you could line break after or and add after each expression.

Yeah no real reason to not merge em.. will do; I think was just in some sort of habit of writing blocks of code that test + return out if some invalid data, condition-by-condition.

… filter-blocks-ui-1

willronchetti · 2020-11-19T21:46:49Z

src/encoded/search/compound_search.py

+    if filter_set.get(CompoundSearchBuilder.TYPE) not in request.registry[TYPES]["FilterSet"].schema["properties"][CompoundSearchBuilder.TYPE]["enum"]:
+        raise HTTPBadRequest("Passed bad {} body param: {}".format(CompoundSearchBuilder.TYPE, filter_set.get(CompoundSearchBuilder.TYPE)))


This is going to cause tests to fail if you don't add the types used in testing to the schema.

willronchetti · 2020-11-19T21:47:09Z

src/encoded/search/search.py

+        # The below didn't work, I guess no request params.. which makes sense..
+        #result.search_base = result.normalize_query(request, result.types, result.doc_types)
+        #result.prepared_terms = result.prepare_search_term(request)
+        #result.set_sort_order()


I think this can be removed now?

… filter-blocks-ui-1

alexkb0009 · 2020-11-20T18:03:48Z

Is deployed now to cgaptest & waiting for indexing to complete.

willronchetti

Approving with small comments - one thing to note is that this may interact weirdly with the new range facets, so may be worth loading up one of those and testing it out. Also needs a version bump in pyproject.toml.

willronchetti · 2020-11-23T16:01:18Z

src/encoded/static/components/item-pages/CaseView/FilteringTab.js

+    let searchHrefBase = "/search/?type=VariantSample";
+    searchHrefBase += initial_search_href_filter_addon ? "&" + initial_search_href_filter_addon : "";
+    searchHrefBase += "&sort=-date_created";


My instinct tells me to watch out for code like this (repeated string building). Strings are still immutable in JS, yes? So why not just do the below? This reduces it from 3 to 2 builds I think. You can reduce it to 1 by including the & in the initial_search_href_filter_addon, which I think would be smart (assuming this isn't optimized later on somehow).

let searchHrefBase = "/search/?type=VariantSample&sort=-date_created"; searchHrefBase += initial_search_href_filter_addon ? "&" + initial_search_href_filter_addon : "";

Ya I might condense this a little.. we do join strings a lot in general in JS tho (i.e. HTML className strings changing re: some state) and this particular case is not really detrimental to performance (it happens once during init for most part, not like 10000 loops of this).

I'd prefer to not bundle in '?' or trailing '&' etc into stored URL chunks since IMO then becomes harder to keep track of what variables are expected to have what prefixes or suffixes; worried that would probably (from personal experience) resort to checking like if not string then ... else if string.endswith("?") or endswith("&") then .. else ... -type stuff which gets even messier lol. Easier/as-easy/cleaner to just do stuff like pathName + (urlQueryChunks ? "?" + urlQueryChunks.join("&") : "") in place(s) where full href is formed/to-be-used.

willronchetti · 2020-11-23T16:04:54Z

src/encoded/static/components/item-pages/CaseView/FilteringTab.js

+    // Load initial filter set Item via AJAX to ensure we get all @@embedded/calculated fields
+    // regardless of how much Case embeds.


This means we can/should be using only default embeds on case then, right?

At least for this active_filterset field, yes... for other things like Pedigree Families I assume the embeds are still needed. We could in future though AJAX in more of the Case rather than embedding it (especially the stuff/data that is click(+) away). Like AJAX in Family (with all its members) and then render the Pedigree visualization.

willronchetti · 2020-11-23T16:18:42Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+         * linkTos must be transformed to UUIDs before POST as well.
+         * Hardcoded here since UI is pretty specific to it.
+         */
+        const fieldsToKeepPrePatch = ["title", "filter_blocks", "search_type", "flags", "created_in_case_accession", "uuid", "status"];


global_flags is the way to quickly pass a sub-query to all filter blocks. For example you might want to apply AF < .001 to all filter blocks - your option is to either go through each filter block and add it individually or set global_flags. Not adding it to the schema is probably an oversight on my part, as global_flags was added in the second iteration. Odds are you'll want to use this functionality (and save it).

willronchetti · 2020-11-23T16:23:52Z

src/encoded/static/components/item-pages/CaseView/FilteringTableFilterSetUI.js

+            <form className="d-flex align-items-center mb-0" action="#case-info.filtering" onSubmit={function(e){
+                e.stopPropagation();
+                e.preventDefault();
+                setIsEditingTitle(false);


Is it convention to use anonymous functions for things like this (and not organize them somehow as named functions)?

Nope... especially not when are passing functions down to named Components since this will cause memoized/PureComponents to re-render each time (since brand new func passed down each time). It becomes less of a concern when is passed into a JSX element (lowercase JSX) (as opposed to React Component (capitalized JSX)) as is here (into "<form ..") since there's little/less memoization optimization taking place there (JSX elements are terminal "leaves" of the React component tree), and especially if the parent Component itself is memoized and prevents unnecessary re-renders.

This anonymous function is the way it is (not named class method or a named memo-generated function) mostly due to the above (the performance benefit is not too significant if formalize it into named func) and due to speed of implementing it. As am polishing it now though, yeah making it a named function would be good/better practice/convention.

willronchetti · 2020-11-23T16:40:11Z

src/encoded/types/case.py

-        "active_filterset.filter_blocks.query",
-        "active_filterset.filter_blocks.flags_applied",
-        "active_filterset.flags",
+        "active_filterset.@id",


This isn't necessary, if it is a linkTo in schema then @id will be included as a default embed.

Good point, I might've been figuring to be explicit in its requirement but I guess we have schema for that also.

… filter-blocks-ui-1

alexkb0009 · 2020-11-24T19:02:38Z

src/encoded/search/compound_search.py

-        return merge_query_strings(qstring1, qstring2)
+        for k, v in dict_with_more_vals.items():
+            if k in dict_to_merge_into:
+                dict_to_merge_into[k] += v


Appears that would get auto-unique'd by urllib.parse.urlencode -- not tested, though.

alexkb0009 · 2020-11-24T19:03:40Z

src/encoded/search/search.py

@@ -142,8 +142,6 @@ def from_search(cls, context, request, search, from_=0, size=10):
        :return:
        """
        result = cls(context, request, skip_bootstrap=True)  # bypass (most of) bootstrap
-        result.from_ = from_


Maybe should have kept this and then done result.search = result.search[from_:from_ + to]here also ... idk.

- Update SPCv - Start with FilterSet UI open.

… filter-blocks-ui-1

Bianca-Morris

Works well. I've already shared my thoughts on look/feel on Slack; looking forward to getting some more feedback on that. I know they'll be excited to test!

Although second thought, maybe double check those Travis logs. Doesn't look like there's a lot of info in there as to why it's failing, so maybe just Travis being flaky? Second pair of eyes can't hurt.

alexkb0009 · 2020-12-01T21:36:14Z

src/encoded/search/search.py

+            # Returns a generator as value of es_result['hits']['hits']
+            # Will be returned directly if self.return_generator is true
+            # or converted to list if meant to be HTTP response.
+            # Theoretical but unnecessary future: Consider allowing to return HTTP Stream of results w. return_generator=true (?)


^ Basically describes functionality of metadata_tsv, don't think any good use case for streaming JSON objects or similar tho.

alexkb0009 added 3 commits October 20, 2020 00:11

Ongoing (requires in-dev SPC commit)

3e61436

Ongoing

5d301e8

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

47a493e

… filter-blocks-ui-1

alexkb0009 requested a review from Bianca-Morris October 22, 2020 19:13

alexkb0009 added 7 commits October 22, 2020 17:15

Ongoing UI, move in responsiveGridState

976cd29

Ongoing - FilterSetController

62be2de

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

68983ee

… filter-blocks-ui-1

Increase case.py active_filterset embed list...

7306feb

TODO: Consider AJAXing it initially? Hmm..

Ongoing

11b074f

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

8e8ddee

… filter-blocks-ui-1

ONGOING - Depends on not-yet-pushed SPC edits

514d363

alexkb0009 requested a review from willronchetti October 30, 2020 14:30

alexkb0009 added 2 commits October 30, 2020 10:32

Ongoing / depends on ongoing SPC edits.

6d3c860

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

dcfb88b

… filter-blocks-ui-1

alexkb0009 requested a review from utku-ozturk October 30, 2020 14:38

willronchetti reviewed Nov 2, 2020

View reviewed changes

alexkb0009 added 8 commits November 6, 2020 12:57

ONGOING | Sort & Related

bc0ea38

Update SPCv 0.0.2.80 -> 0.0.2.82

630646f

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

dda3074

… filter-blocks-ui-1

compound_search > sort

12513c0

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

d092669

… filter-blocks-ui-1

Initial filterblock selection, styling.

a6e5e55

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

e3736df

… filter-blocks-ui-1

Use PackageLockLoader to grab package-lock.json

44d3218

alexkb0009 mentioned this pull request Nov 18, 2020

Code-Splitting Edits #267

Closed

alexkb0009 added 2 commits November 18, 2020 11:30

Revert "Use PackageLockLoader to grab package-lock.json"

c8ad74d

This reverts commit 44d3218.

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

33ec419

… filter-blocks-ui-1

alexkb0009 mentioned this pull request Nov 18, 2020

JS Code-Splitting Edits #268

Merged

Bianca-Morris reviewed Nov 19, 2020

View reviewed changes

Minor Edits

6841ea5

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

c2a68a7

… filter-blocks-ui-1

willronchetti reviewed Nov 19, 2020

View reviewed changes

alexkb0009 added 2 commits November 20, 2020 11:50

SPCv update + moving build_initial_columns

4ec5173

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

67af0b6

… filter-blocks-ui-1

alexkb0009 marked this pull request as ready for review November 20, 2020 16:53

Minor cleanup

614e807

willronchetti approved these changes Nov 23, 2020

View reviewed changes

alexkb0009 added 2 commits November 23, 2020 17:01

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

77d1486

… filter-blocks-ui-1

Limit & compound_search updates

fc41dea

alexkb0009 commented Nov 24, 2020

View reviewed changes

alexkb0009 added 4 commits November 24, 2020 20:43

SPCv Update + PyProj + pass down schemas

3e4423a

Add ES 'search session' or 'preference' param +...

552fa9c

- Update SPCv - Start with FilterSet UI open.

Merge branch 'master' of https://github.com/dbmi-bgm/cgap-portal into…

0facd86

… filter-blocks-ui-1

SPCv Update

4594e7d

Bianca-Morris approved these changes Dec 1, 2020

View reviewed changes

Cleanup

658ddc6

alexkb0009 commented Dec 1, 2020

View reviewed changes

alexkb0009 merged commit b32ce5e into master Dec 1, 2020

alexkb0009 deleted the filter-blocks-ui-1 branch December 1, 2020 21:48


		def build_initial_columns(used_type_schemas):

		columns = OrderedDict()

		if filter_set.get(CompoundSearchBuilder.TYPE) not in request.registry[TYPES]["FilterSet"].schema["properties"][CompoundSearchBuilder.TYPE]["enum"]:
		raise HTTPBadRequest("Passed bad {} body param: {}".format(CompoundSearchBuilder.TYPE, filter_set.get(CompoundSearchBuilder.TYPE)))

		// Load initial filter set Item via AJAX to ensure we get all @@embedded/calculated fields
		// regardless of how much Case embeds.

Filter blocks UI (1) #247

Filter blocks UI (1) #247

Conversation

alexkb0009 commented Oct 22, 2020 • edited

Current Implementation Notes

Next Challenges

Remaining Fixes

Possible ToDos

willronchetti left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 18, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 18, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 18, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 19, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Bianca-Morris left a comment • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 20, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 19, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 19, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 commented Nov 20, 2020

willronchetti left a comment • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 24, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 24, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 Nov 24, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Bianca-Morris left a comment • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexkb0009 commented Oct 22, 2020 •

edited

alexkb0009 Nov 18, 2020 •

edited

alexkb0009 Nov 18, 2020 •

edited

alexkb0009 Nov 18, 2020 •

edited

alexkb0009 Nov 19, 2020 •

edited

Bianca-Morris left a comment •

edited

alexkb0009 Nov 20, 2020 •

edited

alexkb0009 Nov 19, 2020 •

edited

alexkb0009 Nov 19, 2020 •

edited

willronchetti left a comment •

edited

alexkb0009 Nov 24, 2020 •

edited

alexkb0009 Nov 24, 2020 •

edited

alexkb0009 Nov 24, 2020 •

edited

Bianca-Morris left a comment •

edited