def run(self, form):
selection = SelectionSearch(form)
queries = selection.get_article_ids_per_query()
if form.cleaned_data["output_type"] == "application/json+clustermap":
clusters, articles = zip(*get_clusters(queries).items())
cluster_queries = get_cluster_queries(clusters)
image, html = get_clustermap_image(queries)
coords = tuple(clustermap_html_to_coords(html))
return json.dumps({
"coords":
coords,
"image":
b64encode(image).decode("ascii"),
"clusters": [{
"query": q,
"articles": tuple(a)
} for q, a in zip(cluster_queries, articles)]
})
headers, rows = get_clustermap_table(queries)
if form.cleaned_data["output_type"] == "application/spss-sav":
# *sigh*.. this code is fugly.
_headers = {str(h): i for i, h in enumerate(headers)}
return table2sav(
Table(rows=list(rows),
columns=list(map(str, headers)),
columnTypes=[int] * len(headers),
cellfunc=lambda row, col: row[_headers[col]]))
dialect = 'excel'
if form.cleaned_data["output_type"] == "text/csv+tab":
dialect = 'excel-tab'
result = StringIO()
csvf = csv.writer(result, dialect=dialect)
csvf.writerow(list(map(str, headers)))
csvf.writerows(sorted(rows))
if form.cleaned_data[
"output_type"] == "application/json+clustermap+table":
return json.dumps({
"csv": result.getvalue(),
"queries": {q.label: q.query
for q in queries}
})
return result.getvalue()
def run(self, form):
selection = SelectionSearch(form)
queries = selection.get_article_ids_per_query()
if form.cleaned_data["output_type"] == "application/json+clustermap":
clusters, articles = zip(*get_clusters(queries).items())
cluster_queries = get_cluster_queries(clusters)
image, html = get_clustermap_image(queries)
coords = tuple(clustermap_html_to_coords(html))
return json.dumps(
{"coords": coords, "image": b64encode(image).decode("ascii"),
"clusters": [
{"query": q, "articles": tuple(a)}
for q, a in zip(cluster_queries, articles)
]}
)
headers, rows = get_clustermap_table(queries)
if form.cleaned_data["output_type"] == "application/spss-sav":
# *sigh*.. this code is fugly.
_headers = {str(h): i for i, h in enumerate(headers)}
return table2sav(Table(
rows=list(rows),
columns=list(map(str, headers)),
columnTypes=[int]*len(headers),
cellfunc=lambda row, col: row[_headers[col]]
))
dialect = 'excel'
if form.cleaned_data["output_type"] == "text/csv+tab":
dialect = 'excel-tab'
result = StringIO()
csvf = csv.writer(result, dialect=dialect)
csvf.writerow(list(map(str, headers)))
csvf.writerows(sorted(rows))
if form.cleaned_data["output_type"] == "application/json+clustermap+table":
return json.dumps({
"csv": result.getvalue(),
"queries": {q.label: q.query for q in queries}
})
return result.getvalue()
def run(self, form):
self.monitor.update(1, "Executing query..")
selection = SelectionSearch(form)
try:
aggregation, primary, secondary, categories, values = self.get_cache(
)
except NotInCacheError:
narticles = selection.get_count()
self.monitor.update(
10,
"Found {narticles} articles. Aggregating..".format(**locals()))
# Get aggregation
codingjobs = form.cleaned_data["codingjobs"]
primary = form.cleaned_data['primary']
secondary = form.cleaned_data['secondary']
value1 = form.cleaned_data['value1']
value2 = form.cleaned_data['value2']
article_ids = selection.get_article_ids()
# This should probably happen in SelectionForm?
coded_articles = CodedArticle.objects.all()
coded_articles = coded_articles.filter(article__id__in=article_ids)
coded_articles = coded_articles.filter(
codingjob__id__in=codingjobs)
coded_article_ids = set(coded_articles.values_list("id",
flat=True))
for field_name in ("1", "2", "3"):
if not coded_article_ids:
break
schemafield = form.cleaned_data["codingschemafield_{}".format(
field_name)]
schemafield_values = form.cleaned_data[
"codingschemafield_value_{}".format(field_name)]
schemafield_include_descendants = form.cleaned_data[
"codingschemafield_include_descendants_{}".format(
field_name)]
if schemafield and schemafield_values:
code_ids = get_code_filter(
schemafield.codebook, schemafield_values,
schemafield_include_descendants)
coding_values = CodingValue.objects.filter(
coding__coded_article__id__in=coded_article_ids)
coding_values = coding_values.filter(
field__id=schemafield.id)
coding_values = coding_values.filter(intval__in=code_ids)
coded_article_ids &= set(
coding_values.values_list("coding__coded_article__id",
flat=True))
codings = Coding.objects.filter(
coded_article__id__in=coded_article_ids)
terms = selection.get_article_ids_per_query()
orm_aggregate = ORMAggregate(codings, flat=False, terms=terms)
categories = list(filter(None, [primary, secondary]))
values = list(filter(None, [value1, value2]))
aggregation = orm_aggregate.get_aggregate(categories, values)
aggregation = sorted(aggregation, key=to_sortable_tuple)
self.set_cache(
[aggregation, primary, secondary, categories, values])
else:
self.monitor.update(
10, "Found in cache. Rendering..".format(**locals()))
if form.cleaned_data.get("primary_fill_zeroes") and hasattr(
primary, 'interval'):
aggregation = list(
aggregate_es.fill_zeroes(aggregation, primary, secondary))
# Matrices are very annoying to construct in javascript due to missing hashtables. If
# the user requests a table, we thus first convert it to a different format which should
# be easier to render.
if form.cleaned_data["output_type"] == "text/json+aggregation+table":
aggregation = aggregation_to_matrix(aggregation, categories)
if form.cleaned_data["output_type"] == "text/csv":
return aggregation_to_csv(aggregation, categories, values)
self.monitor.update(60, "Serialising..".format(**locals()))
return json.dumps(aggregation,
cls=AggregationEncoder,
check_circular=False)
