Olá! Mesmo seguinto a sugestão na aula de inserir import sys
e sys.path.append("airflowalura")
sigo tendo o seguinte erro:
(venv) user@user:~/Documentos/airflowalura$ python3 operators/twitter_operator.py
Traceback (most recent call last):
File "/home/bruna/Documentos/airflowalura/operators/twitter_operator.py", line 4, in <module>
from hook.twitter_hook import TwitterHook
ModuleNotFoundError: No module named 'hook'
Meu código para o twitter_operator.py:
import sys
sys.path.append("airflowalura")
from airflow.models import BaseOperator, DAG, TaskInstance
from hook.twitter_hook import TwitterHook
import json
from datetime import datetime, timedelta
class TwitterOperator(BaseOperator):
def __init__(self, end_time, start_time, query, **kwargs):
self.end_time = end_time
self.start_time = start_time
super().__init__(**kwargs)
def execute(self, context):
end_time = self.end_time
start_time = self.start_time
query = self.query
with open("extract_twitter.json", "w") as output_file:
for pg in TwitterHook(end_time, start_time, query).run():
json.dump(pg, output_file)
output_file.write("\n")
if __name__ == "__main__":
TIMESTAMP_FORMAT = "%Y-%m-%dT%H:%M:%S.00Z"
end_time = datetime.now().strftime(TIMESTAMP_FORMAT)
start_time = (datetime.now() + timedelta(-1)).date().strftime(TIMESTAMP_FORMAT)
query = "data science"
with DAG(dag_id = "TwitterTest", start_date = datetime.now()) as dag:
to = TwitterOperator(query=query, start_time=start_time, end_time=end_time, task_id="test_run")
ti = TaskInstance(task=to)
to.execute(ti.task_id)
Meu código para o twitter_hook.py:
from airflow.providers.http.hooks.http import HttpHook
import requests
from datetime import datetime, timedelta
import json
class TwitterHook(HttpHook):
def __init__(self, end_time, start_time, query, conn_id=None):
self.end_time = end_time
self.start_time = start_time
self.query = query
self.conn_id = conn_id or "twitter_default"
super().__init__(http_conn_id=self.conn_id)
def create_url(self):
TIMESTAMP_FORMAT = "%Y-%m-%dT%H:%M:%S.00Z"
end_time = self.end_time
start_time = self.start_time
query = "data science"
tweet_fields = "tweet.fields=author_id,conversation_id,created_at,id,in_reply_to_user_id,public_metrics,lang,text"
user_fields = "expansions=author_id&user.fields=id,name,username,created_at"
url_raw = f"{self.base_url}/2/tweets/search/recent?query={query}&{tweet_fields}&{user_fields}&start_time={start_time}&end_time={end_time}"
return url_raw
def connect_to_endpoint(self, url, session):
request = requests.Request("GET", url)
prep = session.prepare_request(request)
self.log.info(f"URL: {url}")
return self.run_and_check(session, prep, {})
def paginate(self, url_raw, session):
lista_json_response = []
#imprimir json
response = self.connect_to_endpoint(url_raw, session)
json_response = response.json()
lista_json_response.append(json_response)
contador = 1
while "next_token" in json_response.get("meta",{}) and contador<100:
next_token = json_response['meta']['next_token']
url = f"{url_raw}&next_token={next_token}"
response = self.connect_to_endpoint(url, session)
json_response = response.json()
lista_json_response.append(json_response)
contador += 1
return lista_json_response
def run(self):
session = self.get_conn()
url_raw = self.create_url()
return self.paginate(url_raw, session)
if __name__ == "__main__":
TIMESTAMP_FORMAT = "%Y-%m-%dT%H:%M:%S.00Z"
end_time = datetime.now().strftime(TIMESTAMP_FORMAT)
start_time = (datetime.now() + timedelta(-1)).date().strftime(TIMESTAMP_FORMAT)
query = "data science"
for pg in TwitterHook(end_time, start_time, query).run():
print(json.dumps(pg, indent=4, sort_keys=True))
E print da minha estrutura de pastas: