Skip to content

Latest commit

 

History

History
369 lines (276 loc) · 10.5 KB

README.md

File metadata and controls

369 lines (276 loc) · 10.5 KB
Raw

Google Ads API Report Fetcher (gaarf)

Python version of Google Ads API Report Fetcher tool a.k.a. gaarf. Please see the full documentation in the root README.

Getting started

Prerequisites

  • Python 3.8+
  • pip installed
  • Google Ads API enabled
  • google-ads.yaml file. Learn how to create one here.

Installation and running

  1. create virtual environment and install the tool
python3 -m venv gaarf
source gaarf/bin/activate
pip install google-ads-api-report-fetcher

install the latest development version with pip install -e git+https://github.com/google/ads-api-report-fetcher.git#egg=google-ads-api-report-fetcher\&subdirectory=py

Versions of the library

  • google-ads-api-report-fetcher[bq] - version with BigQuery support
  • google-ads-api-report-fetcher[pandas] - version with Pandas support
  • google-ads-api-report-fetcher[sqlalchemy] - version with SQLalchemy support
  • google-ads-api-report-fetcher[simulator] - version with support for simulating query results instead of calling Google Ads API.
  • google-ads-api-report-fetcher[sheets] - version with Google Sheets support
  • google-ads-api-report-fetcher[full] - full version
  1. Run the tool with gaarf command:
gaarf <queries> [options]

Documentation on available options see in the root README.md.

Using as a library

Once google-ads-api-report-fetcher is installed you can use it as a library.

Initialize GoogleAdsApiClient to connect to Google Ads API

GoogleAdsApiClient is responsible for connecting to Google Ads API and provides several methods for authentication.

from gaarf import GoogleAdsApiClient


# initialize from local file
client = GoogleAdsApiClient(path_to_config="google-ads.yaml")

# initialize from remote file
client = GoogleAdsApiClient(path_to_config="gs://<PROJECT-ID>/google-ads.yaml")

# initialize from dictionary
google_ads_config_dict = {
    "developer_token": "",
    "client_id": "",
    "client_secret": "",
    "refresh_token": "",
    "client_customer_id": "",
    "use_proto_plus": True
}
client = GoogleAdsApiClient(config_dict=google_ads_config_dict)

initialize AdsReportFetcher to get reports

from gaarf.report_fetcher import AdsReportFetcher

report_fetcher = AdsReportFetcher(client)

# create query text
query_text = "SELECT campaign.id AS campaign_id FROM campaign"

# Execute query and store `campaigns` variable
# specify customer_ids explicitly
customer_ids = ['1', '2']
# or perform mcc expansion for mcc 1234567890
customer_ids = report_fetcher.expand_mcc('1234567890')
campaigns = report_fetcher.fetch(query_text, customer_ids)

# perform mcc expansion when calling `fetch` method
campaigns = report_fetcher.fetch(query_text, '1234567890', auto_expand=True)

Use macros in your queries

parametrized_query_text = """
    SELECT
        campaign.id AS campaign_id
    FROM campaign
    WHERE campaign.status = '{status}'
    """
active_campaigns = report_fetcher.fetch(parametrized_query_text, customer_ids,
                                        {"macro": {
                                            "status": "ENABLED"
                                        }})

Define queries

There are three ways how you can define a query:

  • in a variable
  • in a file
  • in a class (useful when you have complex parametrization and validation)
from gaarf.base_query import BaseQuery
from gaarf.io import reader


# 1. define query as a string an save in a variable
query_string = "SELECT campaign.id FROM campaign"

# 2. define path to a query file and read from it
# path can be local
query_path = "path/to/query.sql"
# or remote
query_path = "gs://PROJECT_ID/path/to/query.sql"

# Instantiate reader
reader_client = reader.FileReader()
# And read from the path
query = reader_client.read(query_path)

# 3. define query as a class

# New style
class Campaigns(BaseQuery):
    query_text  = """
        SELECT
            campaign.id
        FROM campaign
        WHERE campaign.status = {status}
        """

    def __init__(self, status: str = "ENABLED") -> None:
        self.status = status

# Dataclass style
from dataclasses import dataclass

@dataclass
class Campaigns(BaseQuery):
    query_text  = """
        SELECT
            campaign.id
        FROM campaign
        WHERE campaign.status = {status}
        """
    status: str = "ENABLED"

# Old style
class Campaigns(BaseQuery):
    def __init__(self, status: str = "ENABLED"):
        self.query_text = f"""
        SELECT
            campaign.id
        FROM campaign
        WHERE campaign.status = {status}
        """

active_campaigns = report_fetcher.fetch(Campaigns())
inactive_campaigns = report_fetcher.fetch(Campaigns("INACTIVE"))

Iteration and slicing

AdsReportFetcher.fetch method returns an instance of GaarfReport object which you can use to perform simple iteration.

query_text = """
    SELECT
        campaign.id AS campaign_id,
        campaign.name AS campaign_name,
        metrics.clicks AS clicks
    FROM campaign
    WHERE segments.date DURING LAST_7_DAYS
    """
campaigns = report_fetcher.fetch(query_text, '1234567890', auto_expand=True)

# iterate over each row of `campaigns` report
for row in campaigns:
    # Get element as an attribute
    print(row.campaign_id)

    # Get element as a slice
    print(row["campaign_name"])

    # Get element as an index (will print number of clicks)
    print(row[2])

    # Create new column
    row["new_campaign_id"] = row["campaign_id"] + 1

You can easily slice the report

# Create new reports by selecting one or more columns
campaign_only_report = campaigns["campaign_name"]
campaign_name_clicks_report = campaigns[["campaign_name", "clicks"]]

# Get subset of the report
# Get first row only
first_campaign_row = campaigns[0]
# Get first ten rows from the report
first_10_rows_from_campaigns = campaigns[0:10]

Convert report

GaarfReport can be easily converted to common data structures:

# convert `campaigns` to list of lists
campaigns_list = campaigns.to_list()

# convert `campaigns` to flatten list
campaigns_list = campaigns.to_list(row_type="scalar")

# convert `campaigns` column campaign_id to list
campaigns_list = campaigns["campaign_id"].to_list()

# convert `campaigns` column campaign_id to list with unique values
campaigns_list = campaigns["campaign_id"].to_list(distinct=True)

# convert `campaigns` to list of dictionaries
# each dictionary maps report column to its value, i.e.
# {"campaign_name": "test_campaign", "campaign_id": 1, "clicks": 10}
campaigns_list = campaigns.to_list(row_type="dict")

# convert `campaigns` to pandas DataFrame
campaigns_df = campaigns.to_pandas()

# convert `campaigns` to dictionary
# map campaign_id to campaign_name one-to-one
campaigns_df = campaigns.to_dict(
    key_column="campaign_id",
    value_column="campaign_name",
    value_column_output="scalar",
    )

# convert `campaigns` to dictionary
# map campaign_id to campaign_name one-to-many
campaigns_df = campaigns.to_dict(
    key_column="campaign_id",
    value_column="campaign_name",
    value_column_output="list",
    )

Build report

GaarfReport can be easily built from pandas data frame:

import pandas as pd

df = pd.DataFrame(data=[[1]], columns=["one"])
report = GaarfReport.from_pandas(df)

Save report

GaarfReport can be easily saved to local or remote storage:

from gaarf.io import writers

# initialize CSV writer
csv_writer = writers.csv_writer.CsvWriter(destination_folder="/tmp")

# initialize BigQuery writer
bq_writer = writers.bigquery_writer.BigQueryWriter(
    project="", dataset="", location="")

# initialize SQLAlchemy writer
sqlalchemy_writer = writers.sqlalchemy_writer.SqlAlchemyWriter(
    connection_string="")

# initialize Console writer
console_writer = writers.console_writer.ConsoleWriter(page_size=10)

# initialize Json writer
json_writer = writers.json_writer.JsonWriter(destination_folder="/tmp")

# initialize Google Sheets writer
sheet_writer = writers.sheets_writer.SheetWriter(
    share_with="[email protected]",
    credential_files="path/to/credentials.json"
    )


# save report using one of the writers
csv_writer.write(campaigns, destination="my_file_name")
bq_writer.write(campaigns, destination="my_table_name")
sqlalchemy_writer.write(campaigns, destination="my_table_name")
json_writer.write(campaigns, destination="my_table_name")
sheet_writer.write(campaigns, destination="my_table_name")

Combine fetching and saving with AdsQueryExecutor

If your job is to execute query and write it to local/remote storage you can use AdsQueryExecutor to do it easily.

When reading query from file AdsQueryExecutor will use query file name as a name for output file/table.

from gaarf.io import reader, writers
from gaarf.executors import AdsQueryExecutor


# initialize query_executor to fetch report and store them in local/remote storage
query_executor = AdsQueryExecutor(client)

# initialize writer
csv_writer = writers.csv_writer.CsvWriter(destination_folder="/tmp")
reader_client = reader.FileReader()

query_text = """
    SELECT
        campaign.id AS campaign_id,
        campaign.name AS campaign_name,
        metrics.clicks AS clicks
    FROM campaign
    WHERE segments.date DURING LAST_7_DAYS
    """

# execute query and save results to `/tmp/campaign.csv`
query_executor.execute(
    query_text=query_text,
    query_name="campaign",
    customer_ids=customer_ids,
    write_client=csv_writer)

# execute query from file and save to results to `/tmp/query.csv`
query_path="path/to/query.sql"
query_executor.execute(
    query_text=reader_client.read(query_path),
    query_name=query_path,
    customer_ids=customer_ids,
    write_client=csv_writer)

Python specific command line flags

  • --optimize-performance - accepts one of the following values:
    • NONE - no optimizations are done
    • PROTOBUF - convert Google Ads API response to protobuf before parsing (speeds up query execution 5x times but forces conversion of ENUMs to integers instead of strings)
    • BATCH - converts all response of Ads API to a list and then parses its content in parallel
    • BATCH_PROTOBUF - combines BATCH and PROTOBUF approaches.

Disclaimer

This is not an officially supported Google product.