def get_khan_topic_tree(lang="en", curr_key=None):
if lang == "sw":
response = make_request(V2_API_URL.format(lang="swa",
projection=PROJECTION_KEYS),
timeout=120)
else:
response = make_request(V2_API_URL.format(lang=lang,
projection=PROJECTION_KEYS),
timeout=120)
topic_tree = ujson.loads(response.content)
# if name of lang is passed in, get language code
if getlang_by_name(lang):
lang = getlang_by_name(lang).primary_code
if lang not in SUPPORTED_LANGS:
global translations
translations = retrieve_translations(lang_code=lang)
# Flatten node_data
flattened_tree = [
node for node_list in topic_tree.values() for node in node_list
]
# convert to dict with ids as keys
tree_dict = {node["id"]: node for node in flattened_tree}
return _recurse_create(tree_dict["x00000000"], tree_dict, lang=lang)
def generate_dubbed_video_mappings_from_csv():
resp = make_request(KA_LITE_DUBBED_LIST, timeout=120)
csv_data = resp.content.decode("utf-8")
# This CSV file is in standard format: separated by ",", quoted by '"'
reader = csv.reader(StringIO(csv_data))
video_map = {}
header_row = []
# Loop through each row in the spreadsheet.
for row in reader:
# skip over the header rows
if row[0].strip() in ["", "UPDATED:"]:
continue
elif row[0] == "SERIAL":
# Read the header row.
header_row = [
v.lower() for v in row
] # lcase all header row values (including language names)
slug_idx = header_row.index("title id")
english_idx = header_row.index("english")
assert slug_idx != -1, "Video slug column header should be found."
assert english_idx != -1, "English video column header should be found."
else:
# Rows 6 and beyond are data.
assert len(row) == len(
header_row), "Values line length equals headers line length"
# Grab the slug and english video ID.
video_slug = row[slug_idx]
english_video_id = row[english_idx]
assert english_video_id, "English Video ID should not be empty"
assert video_slug, "Slug should not be empty"
# English video is the first video ID column,
# and following columns (until the end) are other languages.
# Loop through those columns and, if a video exists,
# add it to the dictionary.
for idx in range(english_idx, len(row)):
if not row[idx]: # make sure there's a dubbed video
continue
lang = header_row[idx]
if lang not in video_map: # add the first level if it doesn't exist
video_map[lang] = {}
dubbed_youtube_id = row[idx]
if english_video_id == dubbed_youtube_id and lang != "english":
print(
"Removing entry for (%s, %s): dubbed and english youtube ID are the same."
% (lang, english_video_id))
else:
video_map[lang][english_video_id] = row[
idx] # add the corresponding video id for the video, in this language.
return video_map
def get_assessment_items(self):
items_list = []
kalang = ASSESSMENT_LANGUAGE_MAPPING.get(self.lang, self.lang)
for ai_id in self.assessment_items:
item_url = ASSESSMENT_URL.format(assessment_item=ai_id, kalang=kalang)
item = make_request(item_url).json()
# check if assessment item is fully translated, before adding it to list
if item["isFullyTranslated"]:
ai = KhanAssessmentItem(item["id"], item["itemData"], self.source_url)
items_list.append(ai)
return items_list
def get_video_id_english_mappings():
projection = json.dumps(
{"videos": [OrderedDict([("youtubeId", 1), ("id", 1)])]})
r = make_request(V2_API_URL.format(lang='en', projection=projection),
timeout=120)
english_video_data = r.json()
english_video_data = english_video_data["videos"]
mapping = {n["id"]: n["youtubeId"] for n in english_video_data}
return mapping
def get_assessment_items(self):
items_list = []
lang = ASSESSMENT_LANGUAGE_MAPPING.get(self.lang, self.lang)
for i in self.assessment_items:
item_url = ASSESSMENT_URL.format(assessment_item=i["id"],
lang=lang)
item = make_request(item_url).json()
# check if assessment item is fully translated, before adding it to list
if item["is_fully_translated"]:
items_list.append(
KhanAssessmentItem(item["id"], item["item_data"],
self.source_url))
return items_list
def retrieve_translations(lang_code, includes="*.po"):
if lang_code in SUPPORTED_LANGS:
return {}
lang_code = CROWDIN_LANGUAGE_MAPPING.get(lang_code, lang_code)
r = make_request(CROWDIN_URL.format(key=os.environ['KA_CROWDIN_SECRET_KEY'], lang_code=lang_code), timeout=180)
with open('crowdin.zip', "wb") as f:
for chunk in r.iter_content(1024):
f.write(chunk)
zip_extraction_path = tempfile.mkdtemp()
with zipfile.ZipFile('crowdin.zip') as zf:
zf.extractall(zip_extraction_path)
all_filenames = glob.iglob(
os.path.join(zip_extraction_path, "**"),
recursive=True
)
filenames = fnmatch.filter(all_filenames, includes)
# use the polib library, since it's much faster at concatenating
# po files. it doesn't have a dict interface though, so we'll
# reread the file using babel.Catalog.
with tempfile.NamedTemporaryFile() as f:
main_pofile = polib.POFile(fpath=f.name)
for filename in filenames:
pofile = polib.pofile(filename)
main_pofile.merge(pofile)
for entry in main_pofile:
entry.obsolete = False
main_pofile.save()
shutil.rmtree(zip_extraction_path)
msgid_mapping = Catalog(main_pofile)
return msgid_mapping
def get_khan_api_json(lang, update=False):
"""
Get all data for language `lang` from the KA API at /api/v2/topics/topictree
"""
filename = 'khan_academy_json_{}.json'.format(lang)
filepath = os.path.join(KHAN_API_CACHE_DIR, filename)
if os.path.exists(filepath) and not update:
print('Loaded KA API json from cache', filepath)
data = json.load(open(filepath))
else:
print('Downloading KA API json for lang =', lang)
url = V2_API_URL.format(lang=lang, projection=PROJECTION_KEYS)
LOGGER.debug('khan API url=' + url)
response = make_request(url, timeout=120)
data = response.json()
if not os.path.exists(KHAN_API_CACHE_DIR):
os.makedirs(KHAN_API_CACHE_DIR, exist_ok=True)
json.dump(data, open(filepath, 'w'), ensure_ascii=False, indent=4)
return data
def generate_common_core_mapping():
resp = make_request(COMMON_CORE_SPREADSHEET, timeout=120)
csv_data = resp.content.decode("utf-8")
# This CSV file is in standard format: separated by ",", quoted by '"'
reader = csv.reader(StringIO(csv_data))
slug_standard_map = {}
header_row = []
# Loop through each row in the spreadsheet.
for row in reader:
if row[0] == "Grade":
# Read the header row.
header_row = [v.lower()
for v in row] # lcase all header row values
grade_idx = header_row.index("grade")
common_core_idx = header_row.index("common core area")
standard_idx = header_row.index("standard")
skill_name_idx = header_row.index("name of skill on khan academy")
link_idx = header_row.index("link to skill")
description_idx = header_row.index("description")
area_idx = header_row.index("area")
else:
# Grab CC standard and link to exercise
standard_tag = row[standard_idx]
link = row[link_idx]
if not link or not standard_tag:
continue
# parse out slug from link and set standard tag
slug = link.split("e/")[1]
slug_standard_map[slug] = standard_tag
return slug_standard_map
