Skip to content

Latest commit

 

History

History
335 lines (305 loc) · 15.9 KB

README.md

File metadata and controls

335 lines (305 loc) · 15.9 KB
Raw

databridge-etl-tools

Command line tools to extract and load SQL and Carto tables using JSON Table Schema.

Overview

Use this tool to extract data from an Oracle SDE database, load it to S3, and then load it to PostGIS or Carto from S3. In order for this to work, a JSON table schema for the table you are working with needs to reside in S3, in the bucket citygeo-airflow-databridge2 in the schemas/ folder.

The tool can be used either with Docker or as a standalone Python package.

Requirements

(w/o Docker)

  • Python >=3.7, <3.10 due to dependencies in the arcgis package.
  • Pip
  • AWS CLI
  • Oracle 11g Client
  • Postgres
  • Access to citygeo-oracle-instant-client S3 bucket

(w/ Docker)

  • Docker
  • Access to citygeo-oracle-instant-client S3 bucket

Installation

  • Perform the following if installing this package onto a server that has not already had these steps completed
    • Install aws CLI if you don't have it sudo apt-get install awscli
    • Install alien if don't already have it sudo apt-get install alien
    • Run bash ./scripts/pull-oracle-rpm.sh
      • You can then verify with ls that the oracle instant client was downloaded into the project folder
    • Run alien -i oracle-instantclient18.5-basiclite-18.5.0.0.0-3.x86_64.rpm \ && rm oracle-instantclient18.5-basiclite-18.5.0.0.0-3.x86_64.rpm

(with pip)

  • Source a virtual environment, then
pip install git+https://github.com/CityOfPhiladelphia/databridge-etl-tools.git

(w/o Docker)

  • Set environment variables for Oracle, if the below are not already in your ~/.bashrc file then add them there and run them in the terminal as well: 
    export ORACLE_HOME=/usr/lib/oracle/18.5/client64
export LD_LIBRARY_PATH=$ORACLE_HOME/lib
export PATH="$PATH:$ORACLE_HOME"
  • Create a virtual environment if one does not already exist
  • Source your virtual environment
  • On Python3.9 (and maybe other versions) remove the version dependencies on the following packages:
    • pyproj
    • arcgis
  • Install the following necessary packages (note that this is copied from the Dockerfile, which essentially does a similar process) 
    sudo apt-get install --no-install-recommends
    python3-dev \
    libkrb5-dev \
    libsasl2-dev \
    libssl-dev \
    libffi-dev \
    build-essential \
    libblas-dev \
    liblapack-dev \
    libpq-dev \
    netbase \
    apt-utils \
    unzip \
    curl \
    netcat \
    locales \
    git \
    alien \
    libgdal-dev \
    libgeos-dev \
    binutils \
    libproj-dev \
    gdal-bin \
    libspatialindex-dev \
    libaio1 \
    freetds-dev
  • Attempt to install the python3-dev that matches your python minor version, so for example
    • sudo apt-get install python3.9-dev
  • pip install . or pip install -e ./ if you are contributing to the repository
    • Note, you do not need the file docker-fast-requirements.txt

(w/ Docker)

  • Until the dockerfile is able to auto-pull AWS credentials from ~/.aws/credentials, substitute the correct keys and run:
    • export aws_access_key_id=<aws_access_key_id>
    • export aws_secret_access_key=<aws_secret_access_key>
  • Ensure docker is installed on the system - it may be necessary to run sudo apt install docker.io
  • Download the oracle-client (oracle-instantclient18.5-basiclite-18.5.0.0.0-3.x86_64.rpm) to the project folder from the citygeo-oracle-instant-client bucket
  • Run docker build -f Dockerfile.fast -t dbtools --build-arg AWS_ACCESS_KEY_ID="$aws_access_key_id" --build-arg AWS_SECRET_ACCESS_KEY="$aws_secret_access_key" ./

Usage

This package uses a nested series of commands (via sub-modules) to implement separation of concerns. This makes it easier to isolate any bugs and offer additional functionality over time. At any time, add --help to the command to review the help guide for that command or sub-group. In click, commands and sub-groups are internally the same thing.

All commands will take the form of

databridge_etl_tools \
    GROUP \
    GROUP_ARGS \
    COMMAND or SUB_GROUP1 \
    [COMMAND ARGS or SUB_GROUP1 ARGS] \
    [COMMAND or SUB_GROUP2] \
    [COMMAND ARGS or SUB_GROUP2 ARGS] ...

See GROUPS, ARGS, SUB-GROUPS and COMMANDS at the end of this README for a full documentation of all possible methods.

Examples

# Remember to add --help after any command for an explanation and its parameters

# Upsert a record into AGO using a primary key
databridge_etl_tools \
    ago \
    --ago_org_url test \
    --ago_user test \
    --ago_pw test \
    --ago_item_name test \
    --s3_bucket test \
    --s3_key test \
    append-group \
    --in_srid 0 \
    --clean_columns test \
    --batch_size 0 \
    upsert \
    --primary_key test

# Extract a table from Oracle SDE to S3
databridge_etl_tools \
    oracle \
    --table_name li_appeals_type \
    --table_schema gis_lni \
    --connection_string <user>/<password>@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)      (HOST=<host_name>)(PORT=<port>))(CONNECT_DATA=(SID=<dbname>))) \
    --s3_bucket s3-bucket \
    --s3_key s3_key \
    extract

# Load a table from S3 to Carto
databridge_etl_tools \
    carto \
	--table_name test \
	--connection_string carto://user:apikey \
	--s3_bucket test \
	--s3_key test \
	--select_users optional \
	--index_fields optional 
    update

# Load a table from S3 to Postgres
databridge_etl_tools \
    postgres \
	--table_name li_appeals_type \
	--table_schema lni \
	--connection_string postgresql://user:password@host:port/db_name \
	--s3_bucket test \
	--s3_key test \
    load \
	--json_schema_s3_key test

# Extract a json schema from Oracle (right now needed for carto uploads)
databridge_etl_tools \
    oracle \
    --table_name li_appeals_type \
    --table_schema gis_lni \
    --connection_string <user>/<password>@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)      (HOST=<host_name>)(PORT=<port>))(CONNECT_DATA=(SID=<dbname>))) \
    --s3_bucket=test \
    --s3_key=schemas/lni/li_appeals_type.json \
    extract-json-schema

# Extract a json schema from DB2 (needed for carto uploads)
databridge_etl_tools \
    postgres \
    --table_name fema_floodplain_2023 \
    --table_schema viewer_planning \
    --connection_string postgresql://postgres:<password>@<hostname>:5432/<dbname> \
    --s3_bucket=test \
    --s3_key=schemas/planning/fema_floodplain_2023.json \
    extract-json-schema

# Extract from Airtable
databridge_etl_tools \
    airtable \
    --app_id <some_app_id> \
    --api_key <key_for_some_app> \
    --table_name immigrant_services \
    --s3_bucket test \
    --s3_key staging/test/immigrant_services.csv \
    extract

# Extract from Knack
databridge_etl_tools \
    knack \
    --app_id <some_app_id> \
    --api_key <key_for_some_app> \
    --knack_objectid <some_id_corresponding_to_table> \
    --s3_bucket test \
    --s3_key staging/test/candidates.csv \
    extract

Development

To manually test while developing, the package can be entered using the -m module flag (due to the presence of the __main__.py file)

python -m databridge_etl_tools \
    oracle \
    --table_name li_appeals_type \
    --table_schema gis_lni \
    --connection_string <user>/<password>@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=<host_name>)(PORT=<port>))(CONNECT_DATA=(SID=<dbname>))) \
    --s3_bucket s3-bucket \
    --s3_key s3_key \
    extract

Run tests

To one run suite of tests locally, from the root databridge-etl-tools folder, run

pytest tests/<test_database.py>
    --user $USER 
    --password $PASSWORD 
    --host $HOST 
    --database $DATABASE 
    -xv 
    --showlocals 
    --disable-warnings 
    --show-capture stdout

To run just one test, add ::test_name to the end of test_database.py

Deployment

When a commit is pushed to the master branch, GitHub actions will automatically run the tests given in .github/workflows/test_pr_build.yml using the secrets located in the repository in Settings > Secrets and Variables.

For this reason you should make changes to your test branch, make sure they pass automated tests and manual QA testing before merging to master.

GROUPS, ARGS, SUB-GROUPS and COMMANDS:

  • ago: Run ETL commands for AGO
    • Args:
      • --ago_org_url TEXT [required]
      • --ago_user TEXT [required]
      • --ago_pw TEXT [required]
      • --ago_item_name TEXT [required]
      • --s3_bucket TEXT [required]
      • --s3_key TEXT [required]
    • Commands:
      • export Export from an AGO dataset into a csv file in S3
      • post-index-fields Post index fields to AGO
    • Sub-Group:
      • ago-append: Use this group for any commands that utilize append
        • Args:
          • --in_srid INTEGER The SRID of the source datasets geometry features.
          • --clean_columns TEXT Column, or comma separated list of column names to clean of AGO invalid characters.
          • --batch_size INTEGER Size of batch updates to send to AGO
        • Commands:
          • append Appends records to AGO without truncating. NOTE that this is NOT an upsert and will absolutely duplicate rows if you run this multiple times.
          • truncate-append Truncates a dataset in AGO and appends to it from a CSV.
          • upsert Upserts records to AGO, requires a primary key. Upserts the entire CSV into AGO, it does not look for changes or differences.
            • Args: --primary_key TEXT [required]
  • carto: Run ETL commands for Carto:
    • Args:
      • --table_name TEXT [required]
      • --connection_string TEXT [required]
      • --s3_bucket TEXT [required]
      • --s3_key TEXT [required]
      • --select_users TEXT
      • --index_fields TEXT
    • Commands:
      • update Loads a datasets from S3 into carto
  • db2: Run ETL commands for DB2
    • Args:
      • --table_name TEXT [required]
      • --account_name TEXT [required]
    • Commands:
      • copy-staging-to-enterprise Copies from etl_staging to the specified enterprise authoritative dataset.
      • update-oracle-scn WRONG Creates a staging table in etl_staging from the specified enterprise authoritative dataset.
        • Args:
          • --oracle_conn_string TEXT
    • Sub-Group:
      • libpq Use this group for any commands that utilize libpq
        • Args:
          • --enterprise_schema TEXT [required]
          • --libpq_conn_string TEXT [required]
        • Commands:
          • copy-dept-to-enterprise Copy from the dept table directly to an enterpise able in a single transaction that can roll back if it fails.
          • create-staging-from-enterprise Creates a staging table in etl_staging from the specified enterprise authoritative dataset.
  • opendata: Run ETL commands for OpenData
    • Args:
      • --table_name TEXT [required]
      • --table_schema TEXT [required]
      • --s3_bucket TEXT [required]
      • --s3_key TEXT [required]
      • --libpq_conn_string TEXT [required]
      • --opendata_bucket TEXT [required]
    • Commands:
      • upload Takes a CSV from S3, runs some transformations, and then uploads to the specified opendata bucket
  • oracle: Run ETL commands for Oracle
    • Args:
      • --table_name TEXT [required]
      • --table_schema TEXT [required]
      • --connection_string TEXT [required]
      • --s3_bucket TEXT [required]
      • --s3_key TEXT [required]
    • Commands:
      • extract Extracts a dataset in Oracle into a CSV file in S3
      • extract-json-schema Extracts a dataset's schema in Oracle into a JSON file in S3
  • postgres: Run ETL commands for Postgres
    • Args:
      • --table_name TEXT [required]
      • --table_schema TEXT [required]
      • --connection_string TEXT [required]
      • --s3_bucket TEXT
      • --s3_key TEXT
    • Commands:

      • extract Extracts data from a postgres table into a CSV file in S3. Has spatial and SRID detection and will output it in a way that the ago append commands will recognize.

        • Args:
          • --with_srid BOOLEAN Likely only needed for certain views. This controls whether the geopetl frompostgis() function exports with geom_with_srid. That wont work for some views so just export without. [default: True]

      • extract-json-schema Extracts a dataset's schema in Postgres into a JSON file in S3

      • load Prepare and COPY a CSV from S3 to a Postgres table. The keyword arguments "column_mappings" or "mappings_file" can be used to map data file columns to database table colums with different names. Only one of column_mappings or mappings_file should be provided. Note that only the columns whose headers differ between the data file and the database table need to be included. All column names must be quoted.

        • Args:
          • --column_mappings TEXT A string that can be read as a dictionary using ast.literal_eval(). It should take the form "{'data_col': 'db_table_col', 'data_col2': 'db_table_col2', ...}"
          • --mappings_file TEXT A text file that can be opened with open() and that contains one Python dict that can be read with ast.literal_eval(). The file should take the form {"data_col": "db_table_col", "data_col2": "db_table_col2", ... }. Note no quotes around the curly braces {}.

      • upsert-csv Upserts data from a CSV to a Postgres table, which must have at least one primary key. The keyword arguments "column_mappings" or "mappings_file" can be used to map data file columns to database table colums with different names. Only one of column_mappings or mappings_file should be provided. Note that only the columns whose headers differ between the data file and the database table need to be included. All column names must be quoted.

        • Args:
          • --column_mappings TEXT A string that can be read as a dictionary using ast.literal_eval(). It should take the form "{'data_col': 'db_table_col', 'data_col2': 'db_table_col2', ...}"
          • --mappings_file TEXT A text file that can be opened with open() and that contains one Python dict that can be read with ast.literal_eval(). The file should take the form {"data_col": "db_table_col", "data_col2": "db_table_col2", ... }. Note no quotes around the curly braces {}.

      • upsert_table Upserts data from a Postgres table to a Postgres table in the same database, which must have at least one primary key. The keyword arguments "column_mappings" or "mappings_file" can be used to map data file columns to database table colums with different names. Only one of column_mappings or mappings_file should be provided. Note that only the columns whose headers differ between the data file and the database table need to be included. All column names must be quoted.

        • Args:
          • --column_mappings TEXT A string that can be read as a dictionary using ast.literal_eval(). It should take the form "{'data_col': 'db_table_col', 'data_col2': 'db_table_col2', ...}"
          • --mappings_file TEXT A text file that can be opened with open() and that contains one Python dict that can be read with ast.literal_eval(). The file should take the form {"data_col": "db_table_col", "data_col2": "db_table_col2", ... }. Note no quotes around the curly braces {}.
          • --other_schema TEXT Schema of Postgres table to upsert from. If None or absent, assume the same schema as the table being upserted to
          • --other_table TEXT Name of Postgres table to upsert from [required]