petal-etl-composer 0.1.4

Composable ETL Logic Layer

PETaL - Python Extract Transform and Load

A framework to compose custom ETLs on top of a logical floor.

Follow along in the /examples directory to build up arbitrary pipelines.

(This is modeled on Airflow's DAG Operator model for intuitive composition in Python.)

01. Trivial Case

# The most trivial pipeline possible
with Pipeline("01_noop") as dag:  # Define the pipeline context manager, inner block gets scoped to this object
    source = EmptySource("empty_source")  # Define operators
    sink = NoOpSink("no_op_sink")

    source >> sink  # Compose the operators into a DAG

dag.run()

02. Simple Case - Single ETL Step

# A single ETL step
with Pipeline("02_copy_file_to_file") as dag:
    read_logs = FileReader("read_logs", file_path="../data/example_input.txt")
    pattern_filter = RegexFilter("filter_info", pattern="^INFO")
    write_to_file = FileWriter("write_file", file_path="../data/example_output.txt")

    read_logs >> pattern_filter >> write_to_file

dag.run()

03. One Source, Multiple Sinks

    with Pipeline("03_splitting_streams_to_multiple_destinations") as dag:
        read_logs = FileReader("read_logs", file_path="../data/example_input.txt")
        # Fan-out operator that can be used to split the stream into multiple threads
        splitter = Splitter("split")
        
        # One branch will be just a direct copy
        write_to_file_unfiltered = FileWriter("write_unfiltered", file_path="../data/example_output_unfiltered.txt")
        
        # The other branch will be nice and filtered
        pattern_filter = RegexFilter("filter_info", pattern="^INFO")
        write_to_file_filtered = FileWriter("write_filtered", file_path="../data/example_output_filtered.txt")
        

        # Read the logs into the splitter...
        read_logs >> splitter
        
        # ...then read as many branches from the splitter as you want
        splitter >> pattern_filter >> write_to_file_filtered
        splitter >> write_to_file_unfiltered

    dag.run()

Theory

There are 3 types of Operators - Sources, Sinks, and Non-Terminal Operators. Pipelines in Petal are wrappers around arbitrary Directed Acyclic Graphs (DAGs). A Pipeline is constructed from a DAG and has the following invariants:

1. It must have at least 1 Operator satisfying each of the terminal operator types (ie. at least 1 Source and 1 Sink).
2. Operators are directional (data flows in a particular direction within the Operator).
3. Each Operator has a unique ID and a reference to its upstream and downstream Operators.
4. There are no cycles in the graph.