def main(params):
try:
msg = InputMessage.from_params(params)
text = msg.payload.get('text')
return OutputMessage.create().with_type('cortex/Text').with_payload({'text': 'Echo: '+ text}).to_params()
except:
return OutputMessage.create().with_type('cortex/Text').with_payload({'text': 'error: ' + params}).to_params()
def main(params):
try:
cortex_params = InputMessage.from_params(params)
profile_key = "newsAgent/profiles/{}.json".format(
cortex_params.payload.get("profileId"))
# Download users current profile attributes (if it exists ...)
profileAttributes = safe_load_list_of_jsons_from_managed_content(
cortex_params, profile_key)
eprint(profileAttributes)
user_keyword_affinities = determine_user_keyword_affinity_score(
profileAttributes)
# List the users keywords sorted
userKeywordLikes = [{
"keyword": y[0],
"score": y[1]
} for y in sorted(user_keyword_affinities.items(),
key=lambda x: -1 * x[1])]
eprint(userKeywordLikes)
# List the users topics sorted
topics = {
topic["topic"]: topic["keywords"]
for topic in safe_load_list_of_jsons_from_managed_content(
cortex_params, "newsAgent/datasets/topics.json")
}
eprint(topics)
# Determine User Topic Affinities ...
userTopicLikes = [{
"topic":
topic,
"score":
joint_score_of_affinities(user_keyword_affinities, keywords)
} for topic, keywords in topics.items()]
eprint(userTopicLikes)
return OutputMessage.create().with_payload({
"profileId":
cortex_params.payload.get("profileId"),
"profileAttributes": {
"userAffinityTowardsInsightsWithKeywords": userKeywordLikes,
"userAffinityTowardsInsightsAboutTopics": userTopicLikes,
}
}).to_params()
except Exception as e:
return OutputMessage.create().with_payload({
'message': str(e)
}).to_params()
def main(params):
msg = InputMessage.from_params(params)
text = msg.payload.get('text')
return OutputMessage.create().with_type('cortex/Text').with_payload({
'text':
'Got: ' + text
}).to_params()
def main(params):
try:
cortex_params = InputMessage.from_params(params)
# Download profile attributes (if it exists ...)
profileAttributes = safe_load_list_of_jsons_from_managed_content(cortex_params, "newsAgent/profiles/{}.json".format(cortex_params.payload.get("profileId")))
# eprint(profileAttributes)
# Download news articles to present ...
artilces = safe_load_list_of_jsons_from_managed_content(cortex_params, "newsAgent/datasets/newsArticles.json")
# eprint(artilces)
# if the user does not have a profile, return 10 random artilces ...
if not profileAttributes:
scored_articles = [(0, x) for x in random.sample(artilces, 10)]
eprint("No Profile for User...")
# Here on out, we are assuming the user has a profile ...
else:
# Get User Affinity Score Per Keyword
user_keyword_affinities = determine_user_keyword_affinity_score(profileAttributes)
# eprint(user_keyword_affinities)
# Score Each Article Against Users Afinnity Scores & Sort Them...
scored_articles = list(sorted(
[score_article(a, user_keyword_affinities) for a in artilces],
key=lambda x: -1 * x[0]
))
eprint(scored_articles)
# Take the top 3 Articles ...
insights = turn_scored_articles_into_insights(
cortex_params.payload.get("profileId"),
scored_articles[:3]
)
# - [ ] TODO: Download Insights File and Append Newly Generated Insights to it
eprint(insights)
return OutputMessage.create().with_payload({
"profileId": cortex_params.payload.get("profileId"),
"insights": insights
}).to_params()
except Exception as e:
return OutputMessage.create().with_payload({'message': str(e)}).to_params()
def main(params):
# Parse the function params
msg = InputMessage.from_params(params)
# Get text and payload
text = msg.payload.get('text')
# Compute and create output
return OutputMessage.create().with_type('cortex/Text').with_payload({
'text':
text
}).to_params()
def main(params):
# Parse the function params
msg = InputMessage.from_params(params)
# Get input parameters
source = msg.payload.get('source')
category = msg.payload.get('category')
country = msg.payload.get('country')
query = msg.payload.get('query')
# Get properties
batch_size = msg.properties.get('batch_size')
batch_no = msg.properties.get('batch_no')
api_token = msg.properties.get('api_token')
filter_type = msg.properties.get('filter_type')
# get an API response
response = \
filter_news(country, category, source, query, batch_size, batch_no, api_token, filter_type)
# convert response to strings
url, status, results, error_code, error_message = (str(x)
for x in response)
# return the status and results if no error
if error_code == "None":
return OutputMessage.create().with_payload({
'status': status + ": " + url,
'results': results
}).to_params()
# return error code and message if error
else:
return OutputMessage.create().with_payload({
'status':
status + ": " + url,
'error_code':
error_code,
'error_message':
error_message
}).to_params()
def main(params):
try:
cortex_params = InputMessage.from_params(params)
profile_key = "newsAgent/profiles/{}.json".format(
cortex_params.payload.get("profileId"))
# Download users current profile attributes (if it exists ...)
profileAttributes = safe_load_list_of_jsons_from_managed_content(
cortex_params, profile_key)
eprint(profileAttributes)
userKeywordLikes = Counter(
determine_user_keyword_affinity_score(profileAttributes))
eprint(userKeywordLikes)
keywordsOfInsightsLiked = Counter(
list(
itertools.chain.from_iterable([
set(x["content"]["keywords"])
for x in cortex_params.payload.get("insightsLiked", [])
])))
eprint(keywordsOfInsightsLiked)
updatedKeywordLikeCounts = userKeywordLikes + keywordsOfInsightsLiked
new_profile_attributes = recreate_keyword_based_profile_attributes(
cortex_params.payload.get("profileId"), updatedKeywordLikeCounts)
eprint(new_profile_attributes)
safe_save_list_of_jsons_to_managed_content(cortex_params, profile_key,
new_profile_attributes)
return OutputMessage.create().with_payload({
"profileId":
cortex_params.payload.get("profileId")
}).to_params()
except Exception as e:
return OutputMessage.create().with_payload({
'message': str(e)
}).to_params()
def main(params):
msg = InputMessage.from_params(params)
fake = Faker()
name_of_movie = msg.payload.get('name_of_movie')
movie = {
'name_of_movie': name_of_movie,
'lead_1': fake.name(),
'lead_2': fake.name(),
'support_1': fake.name(),
'support_2': fake.name()
}
return OutputMessage.create().with_type('cortex/Text').with_payload({
'movie':
movie
}).to_params()
def main(params):
# Parse the function params
msg = InputMessage.from_params(params)
# Get source and n
source = msg.payload.get('source')
n = msg.payload.get('n')
# get api token from properties
api = msg.properties.get('token')
# get the top headlines using the news api
top_head = news(source.lower(), n, api)
# Compute and create output
return OutputMessage.create().with_payload({'route':'headline','top_head': top_head,'api':api}).to_params()
def main(params):
datasetReferences = get_dataset_refs(params)
parsedArgs = {
NAME_COMPANIES_REF: datasetReferences.get(NAME_COMPANIES_REF),
"api_endpoint": params.get("apiEndpoint")
}
jwt = params.get("token")
datasetStream = read_dataset(parsedArgs, jwt,
parsedArgs[NAME_COMPANIES_REF])
msg = InputMessage.from_params(params)
text = msg.payload.get('text')
return OutputMessage.create().with_payload({
'data': datasetStream
}).to_params()