Fragment is the Ledger API for engineers that move money. Stop wrangling payment tables, debugging balance errors, and hacking together data pipelines. Start shipping the features that make a difference.
Using pip:
pip install fragment-pythonUsing poetry:
poetry add fragment-pythonGet started by instantiating a Client from fragment.sdk.client. You can generate credentials using the Fragment dashboard
from fragment.sdk.client import Client
graphql_client = Client(
client_id="<client id from the dashboard>",
client_secret="<client secret from the dashboard>",
api_url="<api url from the dashboard>",
auth_url="<auth url from the dashboard>",
auth_scope="<auth scope from the dashboard>",
)
async def print_schema():
get_schema_result = await graphql_client.get_schema("<Your schema key here>")
print(get_schema_result.schema_.json())
import asyncio
loop = asyncio.get_event_loop()
loop.run_until_complete(print_schema())Read the Using custom queries section to learn how to use your own GraphQL queries with the SDK.
If you prefer using a synchronous client instead of an async one, then:
from fragment.sync_sdk.client import Client
graphql_client = Client(
client_id="<client id from the dashboard>",
client_secret="<client secret from the dashboard>",
api_url="<api url from the dashboard>",
auth_url="<auth url from the dashboard>",
auth_scope="<auth scope from the dashboard>",
)
get_schema_result = graphql_client.get_schema("<Your schema key here>")
print(get_schema_result.schema_.json())To post a Ledger Entry defined in your Schema:
await graphql_client.add_ledger_entry(
ik="some-ik",
ledger_ik="your-ledger-ik",
type="user_funds_account",
posted="1968-01-01T16:45:00Z",
parameters=dict(
user_id="user-1",
funding_amount="20000",
)
)To read a Ledger Account's balance:
from fragment.sdk.enums import CurrencyCode
from fragment.sdk.input_types import CurrencyMatchInput
await graphql_client.get_ledger_account_balance(
ledger_ik="your-ledger-ik",
path="liabilities/user:user-1/available",
balance_currency=CurrencyMatchInput(code=CurrencyCode.USD),
)While the SDK comes with GraphQL queries out of the box, you may want to customize these queries for your product. In order to do that:
- Define your custom GraphQL queries in a GraphQL file. For example, in
queries/custom-queries.graphql:
query getSchemaName($key: SafeString!) {
schema(schema: { key: $key }) {
key
name
}
}- Run
fragment-python-client-codegento generate the GraphQL SDK client. GraphQL named queries are converted to snake_case to conform to Python's code conventions. Optionally, pass the--syncflag to generate a synchronous client instead of the default async GraphQL client.
fragment-python-client-codegen \
--input-dir libs/fragment/queries/ \
--target-package-name=custom_queries_package \
--output-dir=libs/fragment- Use the client from the generated package in your product! Apart from the custom query methods, this client is functionally identical to
fragment.sdk.client.Client
from .libs.fragment.custom_queries_package.client import Client
graphql_client = Client(
client_id="<client id from the dashboard>",
client_secret="<client secret from the dashboard>",
api_url="<api url from the dashboard>",
auth_url="<auth url from the dashboard>",
auth_scope="<auth scope from the dashboard>",
)
async def print_schema_name():
# Note that getSchemaName is converted to snake_case automatically
response = await graphql_client.get_schema_name("<Your Schema Key>")
print(response.schema_.key)
import asyncio
loop = asyncio.get_event_loop()
loop.run_until_complete(print_schema())