Skip to main content

Variable fal-scripts-path

In order to find scripts in your project, fal uses the fal-scripts-path dbt variable.

Let's consider the following project structure:

.
├── dbt_project.yml
├── models
│   ├── schema.yml
│   └── some_model.sql
└── scripts
   ├── after.py
   ├── before.py
   └── utils
   ├── my_utils.py
   └── process
   └── process_df.py

By default fal-scripts-path is the dbt project directory (where the dbt_project.yml is located).

But it can be changed by setting in a dbt var. An example of setting the variable would be:

name: "fal_test"
version: "1.0.0"
config-version: 2

vars:
fal-scripts-path: "scripts"

Script reference in schema.yml

Use the fal-scripts-path dbt variable as a base directory when associating to fal scripts.

Referencing a script in your schema.yml with the default fal-scripts-path value looks like:

version: 2

models:
- name: some_model
meta:
fal:
scripts:
before:
# searches in `./` because var has default value
- scripts/before.py
post-hook:
- scripts/after.py

But if the fal-scritps-path value is changed to scripts, like specified above, the schema.yml would be:

version: 2

models:
- name: some_model
meta:
fal:
scripts:
before:
# searches in `./scripts/` because of var
- before.py
post-hook:
- after.py

Script importing during runs

For larger scripts or repeated functionality, you may decide to have several Python files with functions to be imported into your fal scripts.

The fal-scripts-path variable refers to the base directory from which you do your imports. Changing fal-scripts-path also changes the base import directory.

For example; importing a script with the default fal-scripts-path value looks like:

# Searching from the top level: include `script` directory in path
import scripts.utils.my_utils as my_utils
from scripts.utils.process.process_df import some_func

Changing the fal-scripts-path value to scripts, like specified above, would require import changes:

# Searching from the `scripts` directory
import utils.my_utils as my_utils
from utils.process.process_df import some_func