This library simplifies the integration to Factiva Analytics API services that delivers premium news content.
The following services are currently implemented.
- auth: Contains tools to handle UserKey authentication and account statistics.
- Snapshots: Allows to run each snapshot creation, monitoring, download and local exploration, in an individual manner. Also allows to run the whole process within a single method.
- Streams: In addition to creating and getting stream details, contains the methods to easily implement a stream listener and push the content to other locations appropriate for high-available setups.
- Taxonomy: Operations that return taxonomies applied to classify news content.
- ArticleFetcher: Gets article's content by unique identifiers (AN), for display purposes only.
To install this library, run the following commands.
$ pip install --upgrade factiva-analytics
Most Factiva Analytics services are implemented in this library. There may be a delay (commonly weeks) when new features are released and their operations are implemented in this package.
Create an AccountInfo instance that contains a summary of the account's basic information and usage statistics.
from factiva.analytics import AccountInfo
u = AccountInfo(
user_key='abcd1234abcd1234abcd1234abcd1234' # Not needed if the ENV variable FACTIVA_USERKEY is set
)
print(u)<'factiva.analytics.AccountInfo'> ├─user_key: <'factiva.analytics.UserKey'> │ ├─key: ****************************1234 │ └─cloud_token: **********************YKB12sJrkHXX ├─account_name: AccName1234 ├─account_type: account_with_contract_limits ├─active_product: DNA ├─max_allowed_extracted_documents: 8,000,000 ├─max_allowed_extractions: 20 ├─currently_running_extractions: 0 ├─total_extracted_documents: 5,493,078 ├─total_extractions: 4 ├─total_stream_instances: 0 ├─total_stream_subscriptions: 0 ├─extractions_list: <NotLoaded> ├─streams_list: <NotLoaded> ├─enabled_company_identifiers: │ ├─[1]: sedol │ ├─[3]: cusip │ ├─[4]: isin │ └─[5]: ticker_exchange ├─remaining_documents: 2,506,922 └─remaining_extractions: 16
Creates an API request that tests the query and returns the number of matching items in the archive.
from factiva.analytics import SnapshotExplain
my_query = "publication_datetime >= '2023-01-01 00:00:00' AND UPPER(source_code) = 'DJDN'"
my_explain = SnapshotExplain(
user_key='abcd1234abcd1234abcd1234abcd1234', # Not needed if the ENV variable FACTIVA_USERKEY is set
query=my_query)
my_explain.process_job() # This operation can take several seconds to complete
print(my_explain)<'factiva.analytics.SnapshotExplain'> ├─user_key: <'factiva.analytics.UserKey'> │ ├─key: ****************************1234 │ └─cloud_token: **********************YKB12sJrkHXX ├─query: <'factiva.analytics.SnapshotExplainQuery'> │ ├─where: publication_datetime >= '2023-01-01 00:00:00' AND UPPER(source_code) = 'DJDN' │ ├─includes: <NotSet> │ ├─excludes: <NotSet> │ ├─include_lists: <NotSet> │ └─exclude_lists: <NotSet> ├─job_response: <'factiva.analytics.SnapshotExplainJobResponse'> │ ├─job_id: 3ee35a80-0406-4f2b-a999-3e4eb5aa94d8 │ ├─job_link: https://api.dowjones...8/_explain │ ├─job_state: JOB_STATE_DONE │ ├─volume_estimate: 2,482,057 │ └─errors: <NoErrors> └─samples: <NotRetrieved>
Create a new snapshot and download to a local repository just require a few lines of code.
from factiva.analytics import SnapshotExtraction
my_query = "publication_datetime >= '2023-01-01 00:00:00' AND UPPER(source_code) = 'DJDN'"
my_snapshot = SnapshotExtraction(
user_key='abcd1234abcd1234abcd1234abcd1234', # Can be ommited if exist as env variable
query=my_query)
my_snapshot.process_job() # This operation can take several minutes to completeAfter the process completes, the output files are stored in a subfolder named as the Extraction Job ID.
In the previous code a new snapshot is created using my_query as selection criteria and user_key for user authentication. After the job is being validated internally, a Snapshot Id is obtained along with the list of files to download. Files are automatically downloaded to a folder named equal to the snapshot ID, and contents are loaded as a Pandas DataFrame to the variable news_articles. This process may take several minutes, but automates the extraction process significantly.