def simple_search(request, template=None):
"""ES-specific simple search view.
This view is for end user searching of the Knowledge Base and
Support Forum. Filtering options are limited to:
* product (`product=firefox`, for example, for only Firefox results)
* document type (`w=2`, for esample, for Support Forum questions only)
"""
# Redirect to old Advanced Search URLs (?a={1,2}) to the new URL.
a = request.GET.get("a")
if a in ["1", "2"]:
new_url = reverse("search.advanced") + "?" + request.GET.urlencode()
return HttpResponseRedirect(new_url)
# JSON-specific variables
is_json = request.GET.get("format") == "json"
callback = request.GET.get("callback", "").strip()
content_type = "application/x-javascript" if callback else "application/json"
# Check callback is valid
if is_json and callback and not jsonp_is_valid(callback):
return HttpResponse(
json.dumps({"error": _("Invalid callback function.")}), content_type=content_type, status=400
)
language = locale_or_default(request.GET.get("language", request.LANGUAGE_CODE))
r = request.GET.copy()
# TODO: Do we really need to add this to the URL if it isn't already there?
r["w"] = r.get("w", constants.WHERE_BASIC)
# TODO: Break out a separate simple search form.
search_form = SimpleSearchForm(r, auto_id=False)
if not search_form.is_valid():
if is_json:
return HttpResponse(json.dumps({"error": _("Invalid search data.")}), content_type=content_type, status=400)
t = template if request.MOBILE else "search/form.html"
search_ = render(request, t, {"advanced": False, "request": request, "search_form": search_form})
cache_period = settings.SEARCH_CACHE_PERIOD
search_["Cache-Control"] = "max-age=%s" % (cache_period * 60)
search_["Expires"] = (datetime.utcnow() + timedelta(minutes=cache_period)).strftime(EXPIRES_FMT)
return search_
cleaned = search_form.cleaned_data
# On mobile, we default to just wiki results.
if request.MOBILE and cleaned["w"] == constants.WHERE_BASIC:
cleaned["w"] = constants.WHERE_WIKI
page = max(smart_int(request.GET.get("page")), 1)
offset = (page - 1) * settings.SEARCH_RESULTS_PER_PAGE
lang = language.lower()
if settings.LANGUAGES_DICT.get(lang):
lang_name = settings.LANGUAGES_DICT[lang]
else:
lang_name = ""
# We use a regular S here because we want to search across
# multiple doctypes.
searcher = AnalyzerS().es(urls=settings.ES_URLS).indexes(es_utils.read_index("default"))
wiki_f = F(model="wiki_document")
question_f = F(model="questions_question")
cleaned_q = cleaned["q"]
products = cleaned["product"]
if not products and "all_products" not in request.GET:
lowered_q = cleaned_q.lower()
if "thunderbird" in lowered_q:
products.append("thunderbird")
elif "android" in lowered_q:
products.append("mobile")
elif "ios" in lowered_q or "ipad" in lowered_q or "ipod" in lowered_q or "iphone" in lowered_q:
products.append("ios")
elif "firefox os" in lowered_q:
products.append("firefox-os")
elif "firefox" in lowered_q:
products.append("firefox")
# Start - wiki filters
if cleaned["w"] & constants.WHERE_WIKI:
# Category filter
wiki_f &= F(document_category__in=settings.SEARCH_DEFAULT_CATEGORIES)
# Locale filter
wiki_f &= F(document_locale=language)
# Product filter
for p in products:
wiki_f &= F(product=p)
# Archived bit
wiki_f &= F(document_is_archived=False)
# End - wiki filters
# Start - support questions filters
if cleaned["w"] & constants.WHERE_SUPPORT:
# Has helpful answers is set by default if using basic search
cleaned["has_helpful"] = constants.TERNARY_YES
# No archived questions in default search.
cleaned["is_archived"] = constants.TERNARY_NO
# These filters are ternary, they can be either YES, NO, or OFF
ternary_filters = ("has_helpful", "is_archived")
d = dict(
("question_%s" % filter_name, _ternary_filter(cleaned[filter_name]))
for filter_name in ternary_filters
if cleaned[filter_name]
)
if d:
question_f &= F(**d)
# Product filter
for p in products:
question_f &= F(product=p)
# End - support questions filters
# Done with all the filtery stuff--time to generate results
# Combine all the filters and add to the searcher
doctypes = []
final_filter = F()
if cleaned["w"] & constants.WHERE_WIKI:
doctypes.append(DocumentMappingType.get_mapping_type_name())
final_filter |= wiki_f
if cleaned["w"] & constants.WHERE_SUPPORT:
doctypes.append(QuestionMappingType.get_mapping_type_name())
final_filter |= question_f
searcher = searcher.doctypes(*doctypes)
searcher = searcher.filter(final_filter)
if "explain" in request.GET and request.GET["explain"] == "1":
searcher = searcher.explain()
documents = ComposedList()
try:
# Set up the highlights. Show the entire field highlighted.
searcher = searcher.highlight(
"question_content", # support forum
"document_summary", # kb
pre_tags=["<b>"],
post_tags=["</b>"],
number_of_fragments=0,
)
# Set up boosts
searcher = searcher.boost(
question_title=4.0,
question_content=3.0,
question_answer_content=3.0,
document_title=6.0,
document_content=1.0,
document_keywords=8.0,
document_summary=2.0,
# Text phrases in document titles and content get an extra
# boost.
document_title__match_phrase=10.0,
document_content__match_phrase=8.0,
)
# Build the query
query_fields = chain(*[cls.get_query_fields() for cls in [DocumentMappingType, QuestionMappingType]])
query = {}
# Create match and match_phrase queries for every field
# we want to search.
for field in query_fields:
for query_type in ["match", "match_phrase"]:
query["%s__%s" % (field, query_type)] = cleaned_q
# Transform the query to use locale aware analyzers.
query = es_utils.es_query_with_analyzer(query, language)
searcher = searcher.query(should=True, **query)
num_results = min(searcher.count(), settings.SEARCH_MAX_RESULTS)
# TODO - Can ditch the ComposedList here, but we need
# something that paginate can use to figure out the paging.
documents = ComposedList()
documents.set_count(("results", searcher), num_results)
results_per_page = settings.SEARCH_RESULTS_PER_PAGE
pages = paginate(request, documents, results_per_page)
# If we know there aren't any results, let's cheat and in
# doing that, not hit ES again.
if num_results == 0:
searcher = []
else:
# Get the documents we want to show and add them to
# docs_for_page
documents = documents[offset : offset + results_per_page]
if len(documents) == 0:
# If the user requested a page that's beyond the
# pagination, then documents is an empty list and
# there are no results to show.
searcher = []
else:
bounds = documents[0][1]
searcher = searcher[bounds[0] : bounds[1]]
results = []
for i, doc in enumerate(searcher):
rank = i + offset
if doc["model"] == "wiki_document":
summary = _build_es_excerpt(doc)
if not summary:
summary = doc["document_summary"]
result = {"title": doc["document_title"], "type": "document"}
elif doc["model"] == "questions_question":
summary = _build_es_excerpt(doc)
if not summary:
# We're excerpting only question_content, so if
# the query matched question_title or
# question_answer_content, then there won't be any
# question_content excerpts. In that case, just
# show the question--but only the first 500
# characters.
summary = bleach.clean(doc["question_content"], strip=True)[:500]
result = {
"title": doc["question_title"],
"type": "question",
"is_solved": doc["question_is_solved"],
"num_answers": doc["question_num_answers"],
"num_votes": doc["question_num_votes"],
"num_votes_past_week": doc["question_num_votes_past_week"],
}
result["url"] = doc["url"]
result["object"] = doc
result["search_summary"] = summary
result["rank"] = rank
result["score"] = doc.es_meta.score
result["explanation"] = escape(format_explanation(doc.es_meta.explanation))
result["id"] = doc["id"]
results.append(result)
except ES_EXCEPTIONS as exc:
# Handle timeout and all those other transient errors with a
# "Search Unavailable" rather than a Django error page.
if is_json:
return HttpResponse(json.dumps({"error": _("Search Unavailable")}), content_type=content_type, status=503)
# Cheating here: Convert from 'Timeout()' to 'timeout' so
# we have less code, but still have good stats.
exc_bucket = repr(exc).lower().strip("()")
statsd.incr("search.esunified.{0}".format(exc_bucket))
log.exception(exc)
t = "search/mobile/down.html" if request.MOBILE else "search/down.html"
return render(request, t, {"q": cleaned["q"]}, status=503)
items = [(k, v) for k in search_form.fields for v in r.getlist(k) if v and k != "a"]
items.append(("a", "2"))
fallback_results = None
if num_results == 0:
fallback_results = _fallback_results(language, cleaned["product"])
product = Product.objects.filter(slug__in=cleaned["product"])
if product:
product_titles = [_(p.title, "DB: products.Product.title") for p in product]
else:
product_titles = [_("All Products")]
product_titles = ", ".join(product_titles)
data = {
"num_results": num_results,
"results": results,
"fallback_results": fallback_results,
"product_titles": product_titles,
"q": cleaned["q"],
"w": cleaned["w"],
"lang_name": lang_name,
}
if is_json:
# Models are not json serializable.
for r in data["results"]:
del r["object"]
data["total"] = len(data["results"])
data["products"] = [{"slug": p.slug, "title": p.title} for p in Product.objects.filter(visible=True)]
if product:
data["product"] = product[0].slug
pages = Paginator(pages)
data["pagination"] = dict(
number=pages.pager.number,
num_pages=pages.pager.paginator.num_pages,
has_next=pages.pager.has_next(),
has_previous=pages.pager.has_previous(),
max=pages.max,
span=pages.span,
dotted_upper=pages.pager.dotted_upper,
dotted_lower=pages.pager.dotted_lower,
page_range=pages.pager.page_range,
url=pages.pager.url,
)
if not results:
data["message"] = _("No pages matched the search criteria")
json_data = json.dumps(data)
if callback:
json_data = callback + "(" + json_data + ");"
return HttpResponse(json_data, content_type=content_type)
data.update(
{
"product": product,
"products": Product.objects.filter(visible=True),
"pages": pages,
"search_form": search_form,
"advanced": False,
}
)
results_ = render(request, template, data)
cache_period = settings.SEARCH_CACHE_PERIOD
results_["Cache-Control"] = "max-age=%s" % (cache_period * 60)
results_["Expires"] = (datetime.utcnow() + timedelta(minutes=cache_period)).strftime(EXPIRES_FMT)
results_.set_cookie(settings.LAST_SEARCH_COOKIE, urlquote(cleaned["q"]), max_age=3600, secure=False, httponly=False)
return results_
def advanced_search(request, template=None):
"""ES-specific Advanced search view"""
# JSON-specific variables
is_json = request.GET.get("format") == "json"
callback = request.GET.get("callback", "").strip()
content_type = "application/x-javascript" if callback else "application/json"
# Check callback is valid
if is_json and callback and not jsonp_is_valid(callback):
return HttpResponse(
json.dumps({"error": _("Invalid callback function.")}), content_type=content_type, status=400
)
language = locale_or_default(request.GET.get("language", request.LANGUAGE_CODE))
r = request.GET.copy()
# TODO: Figure out how to get rid of 'a' and do it.
# It basically is used to switch between showing the form or results.
a = request.GET.get("a", "2")
# TODO: This is so the 'a=1' stays in the URL for pagination.
r["a"] = 1
# Search default values
try:
category = map(int, r.getlist("category")) or settings.SEARCH_DEFAULT_CATEGORIES
except ValueError:
category = settings.SEARCH_DEFAULT_CATEGORIES
r.setlist("category", category)
r["language"] = language
search_form = AdvancedSearchForm(r, auto_id=False)
search_form.set_allowed_forums(request.user)
# This is all we use a for now I think.
if not search_form.is_valid() or a == "2":
if is_json:
return HttpResponse(json.dumps({"error": _("Invalid search data.")}), content_type=content_type, status=400)
t = template if request.MOBILE else "search/form.html"
search_ = render(request, t, {"advanced": True, "request": request, "search_form": search_form})
cache_period = settings.SEARCH_CACHE_PERIOD
search_["Cache-Control"] = "max-age=%s" % (cache_period * 60)
search_["Expires"] = (datetime.utcnow() + timedelta(minutes=cache_period)).strftime(EXPIRES_FMT)
return search_
cleaned = search_form.cleaned_data
if request.MOBILE and cleaned["w"] == constants.WHERE_BASIC:
cleaned["w"] = constants.WHERE_WIKI
page = max(smart_int(request.GET.get("page")), 1)
offset = (page - 1) * settings.SEARCH_RESULTS_PER_PAGE
lang = language.lower()
if settings.LANGUAGES_DICT.get(lang):
lang_name = settings.LANGUAGES_DICT[lang]
else:
lang_name = ""
# We use a regular S here because we want to search across
# multiple doctypes.
searcher = AnalyzerS().es(urls=settings.ES_URLS).indexes(es_utils.read_index("default"))
wiki_f = F(model="wiki_document")
question_f = F(model="questions_question")
discussion_f = F(model="forums_thread")
# Start - wiki filters
if cleaned["w"] & constants.WHERE_WIKI:
# Category filter
if cleaned["category"]:
wiki_f &= F(document_category__in=cleaned["category"])
# Locale filter
wiki_f &= F(document_locale=language)
# Product filter
products = cleaned["product"]
for p in products:
wiki_f &= F(product=p)
# Topics filter
topics = cleaned["topics"]
for t in topics:
wiki_f &= F(topic=t)
# Archived bit
if not cleaned["include_archived"]:
wiki_f &= F(document_is_archived=False)
# End - wiki filters
# Start - support questions filters
if cleaned["w"] & constants.WHERE_SUPPORT:
# These filters are ternary, they can be either YES, NO, or OFF
ternary_filters = ("is_locked", "is_solved", "has_answers", "has_helpful", "is_archived")
d = dict(
("question_%s" % filter_name, _ternary_filter(cleaned[filter_name]))
for filter_name in ternary_filters
if cleaned[filter_name]
)
if d:
question_f &= F(**d)
if cleaned["asked_by"]:
question_f &= F(question_creator=cleaned["asked_by"])
if cleaned["answered_by"]:
question_f &= F(question_answer_creator=cleaned["answered_by"])
q_tags = [t.strip() for t in cleaned["q_tags"].split(",")]
for t in q_tags:
if t:
question_f &= F(question_tag=t)
# Product filter
products = cleaned["product"]
for p in products:
question_f &= F(product=p)
# Topics filter
topics = cleaned["topics"]
for t in topics:
question_f &= F(topic=t)
# End - support questions filters
# Start - discussion forum filters
if cleaned["w"] & constants.WHERE_DISCUSSION:
if cleaned["author"]:
discussion_f &= F(post_author_ord=cleaned["author"])
if cleaned["thread_type"]:
if constants.DISCUSSION_STICKY in cleaned["thread_type"]:
discussion_f &= F(post_is_sticky=1)
if constants.DISCUSSION_LOCKED in cleaned["thread_type"]:
discussion_f &= F(post_is_locked=1)
valid_forum_ids = [f.id for f in Forum.authorized_forums_for_user(request.user)]
forum_ids = None
if cleaned["forum"]:
forum_ids = [f for f in cleaned["forum"] if f in valid_forum_ids]
# If we removed all the forums they wanted to look at or if
# they didn't specify, then we filter on the list of all
# forums they're authorized to look at.
if not forum_ids:
forum_ids = valid_forum_ids
discussion_f &= F(post_forum_id__in=forum_ids)
# End - discussion forum filters
# Created filter
unix_now = int(time.time())
interval_filters = (
("created", cleaned["created"], cleaned["created_date"]),
("updated", cleaned["updated"], cleaned["updated_date"]),
)
for filter_name, filter_option, filter_date in interval_filters:
if filter_option == constants.INTERVAL_BEFORE:
before = {filter_name + "__gte": 0, filter_name + "__lte": max(filter_date, 0)}
discussion_f &= F(**before)
question_f &= F(**before)
elif filter_option == constants.INTERVAL_AFTER:
after = {filter_name + "__gte": min(filter_date, unix_now), filter_name + "__lte": unix_now}
discussion_f &= F(**after)
question_f &= F(**after)
# Note: num_voted (with a d) is a different field than num_votes
# (with an s). The former is a dropdown and the latter is an
# integer value.
if cleaned["num_voted"] == constants.INTERVAL_BEFORE:
question_f &= F(question_num_votes__lte=max(cleaned["num_votes"], 0))
elif cleaned["num_voted"] == constants.INTERVAL_AFTER:
question_f &= F(question_num_votes__gte=cleaned["num_votes"])
# Done with all the filtery stuff--time to generate results
# Combine all the filters and add to the searcher
doctypes = []
final_filter = F()
if cleaned["w"] & constants.WHERE_WIKI:
doctypes.append(DocumentMappingType.get_mapping_type_name())
final_filter |= wiki_f
if cleaned["w"] & constants.WHERE_SUPPORT:
doctypes.append(QuestionMappingType.get_mapping_type_name())
final_filter |= question_f
if cleaned["w"] & constants.WHERE_DISCUSSION:
doctypes.append(ThreadMappingType.get_mapping_type_name())
final_filter |= discussion_f
searcher = searcher.doctypes(*doctypes)
searcher = searcher.filter(final_filter)
if "explain" in request.GET and request.GET["explain"] == "1":
searcher = searcher.explain()
documents = ComposedList()
try:
cleaned_q = cleaned["q"]
# Set up the highlights. Show the entire field highlighted.
searcher = searcher.highlight(
"question_content", # support forum
"document_summary", # kb
"post_content", # contributor forum
pre_tags=["<b>"],
post_tags=["</b>"],
number_of_fragments=0,
)
# Set up boosts
searcher = searcher.boost(
question_title=4.0,
question_content=3.0,
question_answer_content=3.0,
post_title=2.0,
post_content=1.0,
document_title=6.0,
document_content=1.0,
document_keywords=8.0,
document_summary=2.0,
# Text phrases in document titles and content get an extra
# boost.
document_title__match_phrase=10.0,
document_content__match_phrase=8.0,
)
# Apply sortby for advanced search of questions
if cleaned["w"] == constants.WHERE_SUPPORT:
sortby = cleaned["sortby"]
try:
searcher = searcher.order_by(*constants.SORT_QUESTIONS[sortby])
except IndexError:
# Skip index errors because they imply the user is
# sending us sortby values that aren't valid.
pass
# Apply sortby for advanced search of kb documents
if cleaned["w"] == constants.WHERE_WIKI:
sortby = cleaned["sortby_documents"]
try:
searcher = searcher.order_by(*constants.SORT_DOCUMENTS[sortby])
except IndexError:
# Skip index errors because they imply the user is
# sending us sortby values that aren't valid.
pass
# Build the query
if cleaned_q:
query_fields = chain(
*[cls.get_query_fields() for cls in [DocumentMappingType, ThreadMappingType, QuestionMappingType]]
)
query = {}
# Create a simple_query_search query for every field
# we want to search.
for field in query_fields:
query["%s__sqs" % field] = cleaned_q
# Transform the query to use locale aware analyzers.
query = es_utils.es_query_with_analyzer(query, language)
searcher = searcher.query(should=True, **query)
num_results = min(searcher.count(), settings.SEARCH_MAX_RESULTS)
# TODO - Can ditch the ComposedList here, but we need
# something that paginate can use to figure out the paging.
documents = ComposedList()
documents.set_count(("results", searcher), num_results)
results_per_page = settings.SEARCH_RESULTS_PER_PAGE
pages = paginate(request, documents, results_per_page)
# If we know there aren't any results, let's cheat and in
# doing that, not hit ES again.
if num_results == 0:
searcher = []
else:
# Get the documents we want to show and add them to
# docs_for_page
documents = documents[offset : offset + results_per_page]
if len(documents) == 0:
# If the user requested a page that's beyond the
# pagination, then documents is an empty list and
# there are no results to show.
searcher = []
else:
bounds = documents[0][1]
searcher = searcher[bounds[0] : bounds[1]]
results = []
for i, doc in enumerate(searcher):
rank = i + offset
if doc["model"] == "wiki_document":
summary = _build_es_excerpt(doc)
if not summary:
summary = doc["document_summary"]
result = {"title": doc["document_title"], "type": "document"}
elif doc["model"] == "questions_question":
summary = _build_es_excerpt(doc)
if not summary:
# We're excerpting only question_content, so if
# the query matched question_title or
# question_answer_content, then there won't be any
# question_content excerpts. In that case, just
# show the question--but only the first 500
# characters.
summary = bleach.clean(doc["question_content"], strip=True)[:500]
result = {
"title": doc["question_title"],
"type": "question",
"is_solved": doc["question_is_solved"],
"num_answers": doc["question_num_answers"],
"num_votes": doc["question_num_votes"],
"num_votes_past_week": doc["question_num_votes_past_week"],
}
else:
summary = _build_es_excerpt(doc, first_only=True)
result = {"title": doc["post_title"], "type": "thread"}
result["url"] = doc["url"]
result["object"] = doc
result["search_summary"] = summary
result["rank"] = rank
result["score"] = doc.es_meta.score
result["explanation"] = escape(format_explanation(doc.es_meta.explanation))
results.append(result)
except ES_EXCEPTIONS as exc:
# Handle timeout and all those other transient errors with a
# "Search Unavailable" rather than a Django error page.
if is_json:
return HttpResponse(json.dumps({"error": _("Search Unavailable")}), content_type=content_type, status=503)
# Cheating here: Convert from 'Timeout()' to 'timeout' so
# we have less code, but still have good stats.
exc_bucket = repr(exc).lower().strip("()")
statsd.incr("search.esunified.{0}".format(exc_bucket))
log.exception(exc)
t = "search/mobile/down.html" if request.MOBILE else "search/down.html"
return render(request, t, {"q": cleaned["q"]}, status=503)
items = [(k, v) for k in search_form.fields for v in r.getlist(k) if v and k != "a"]
items.append(("a", "2"))
fallback_results = None
if num_results == 0:
fallback_results = _fallback_results(language, cleaned["product"])
product = Product.objects.filter(slug__in=cleaned["product"])
if product:
product_titles = [_(p.title, "DB: products.Product.title") for p in product]
else:
product_titles = [_("All Products")]
product_titles = ", ".join(product_titles)
data = {
"num_results": num_results,
"results": results,
"fallback_results": fallback_results,
"product_titles": product_titles,
"q": cleaned["q"],
"w": cleaned["w"],
"lang_name": lang_name,
"advanced": True,
}
if is_json:
# Models are not json serializable.
for r in data["results"]:
del r["object"]
data["total"] = len(data["results"])
data["products"] = [{"slug": p.slug, "title": p.title} for p in Product.objects.filter(visible=True)]
if product:
data["product"] = product[0].slug
pages = Paginator(pages)
data["pagination"] = dict(
number=pages.pager.number,
num_pages=pages.pager.paginator.num_pages,
has_next=pages.pager.has_next(),
has_previous=pages.pager.has_previous(),
max=pages.max,
span=pages.span,
dotted_upper=pages.pager.dotted_upper,
dotted_lower=pages.pager.dotted_lower,
page_range=pages.pager.page_range,
url=pages.pager.url,
)
if not results:
data["message"] = _("No pages matched the search criteria")
json_data = json.dumps(data)
if callback:
json_data = callback + "(" + json_data + ");"
return HttpResponse(json_data, content_type=content_type)
data.update(
{
"product": product,
"products": Product.objects.filter(visible=True),
"pages": pages,
"search_form": search_form,
}
)
results_ = render(request, template, data)
cache_period = settings.SEARCH_CACHE_PERIOD
results_["Cache-Control"] = "max-age=%s" % (cache_period * 60)
results_["Expires"] = (datetime.utcnow() + timedelta(minutes=cache_period)).strftime(EXPIRES_FMT)
results_.set_cookie(settings.LAST_SEARCH_COOKIE, urlquote(cleaned["q"]), max_age=3600, secure=False, httponly=False)
return results_