Shaped CLI
CLI for interactions with the Shaped API.
For full documentation, see docs.shaped.ai.
Installing the Shaped CLI
pip install shaped
Initialize
shaped init --api-key <API_KEY> [--env <ENV>]
The --env option defaults to prod and can be set to other environments like dev or staging.
Engine API
Create Engine
shaped create-engine --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped create-engine
Update Engine
shaped update-engine --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped update-engine
List Engines
shaped list-engines
View Engine
shaped view-engine --engine-name <ENGINE_NAME>
Delete Engine
shaped delete-engine --engine-name <ENGINE_NAME>
Table API
Create Table
shaped create-table --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped create-table
Create Table from URI
Create a table and automatically insert data from a file:
shaped create-table-from-uri --name <TABLE_NAME> --path <PATH_TO_FILE> --type <FILE_TYPE>
Supported file types: parquet, csv, tsv, json, jsonl
List Tables
shaped list-tables
Table Insert
Insert data into an existing table:
shaped table-insert --table-name <TABLE_NAME> --file <DATAFRAME_FILE> --type <FILE_TYPE>
Supported file types: parquet, csv, tsv, json, jsonl
View Table
shaped view-table --table-name <TABLE_NAME>
Update Table
shaped update-table --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped update-table
Delete Table
shaped delete-table --table-name <TABLE_NAME>
View API
Create View
shaped create-view --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped create-view
List Views
shaped list-views
View View
shaped view-view --view-name <VIEW_NAME>
Update View
shaped update-view --file <PATH_TO_FILE>
Or pipe from stdin:
cat <PATH_TO_FILE> | shaped update-view
Delete View
shaped delete-view --view-name <VIEW_NAME>
Query API
Execute Query
Execute an ad-hoc query against an engine. The query can be provided in three ways:
From a file:
shaped query --engine-name <ENGINE_NAME> --query-file <PATH_TO_FILE>
As a JSON string:
shaped query --engine-name <ENGINE_NAME> --query '{"sql": "SELECT * FROM table"}'
From stdin:
cat <PATH_TO_FILE> | shaped query --engine-name <ENGINE_NAME>
Execute Saved Query
Execute a previously saved query by name:
shaped execute-saved-query --engine-name <ENGINE_NAME> --query-name <QUERY_NAME>
List Saved Queries
List all saved queries for an engine:
shaped list-saved-queries --engine-name <ENGINE_NAME>
View Saved Query
View the definition of a saved query:
shaped view-saved-query --engine-name <ENGINE_NAME> --query-name <QUERY_NAME>
Python SDK (V2)
Installation
Pip Installation
pip install shaped
V2 Query API with Fluent Builders
The V2 SDK provides a fluent query builder API that leverages Shaped's declarative query language:
from shaped import (
Client,
RankQueryBuilder,
ColumnOrder,
TextSearch,
Ensemble,
Diversity,
)
client = Client(api_key="your-api-key")
# Build a query using the fluent builder with step objects
query = (RankQueryBuilder()
.from_entity('item')
.retrieve(
ColumnOrder(
[{'name': 'popularity', 'ascending': False}],
limit=1000
),
TextSearch(
'laptop',
mode={'type': 'vector', 'text_embedding_ref': 'text_emb'}
)
)
.filter(Expression('price < 1000')) filter
.score(Ensemble('lightgbm', input_user_id='$parameters.userId'))
.reorder(Diversity(diversity_attributes=['category']))
.limit(50)
.columns(['item_id', 'title', 'price'])
.build())
# Execute the query
result = client.execute_query(
'my_engine',
query,
parameters={'userId': '123'}
)
Available Step Factory Functions
The SDK provides convenient factory functions for creating step objects:
Retrieve Steps:
ColumnOrder(columns, limit=100, where=None, name=None)- Sort by columnsTextSearch(input_text_query, mode, limit=100, where=None, name=None)- Text searchSimilarity(embedding_ref, query_encoder, limit=100, where=None, name=None)- Similarity searchFilter(where=None, limit=100, name=None)- Filter without orderingCandidateIds(item_ids, limit=None, name=None)- Specific item IDsCandidateAttributes(item_attributes, limit=None, name=None)- Item attributes
Filter Steps:
Expression(expression, name=None)- DuckDB filter expressionTruncate(max_length=500, name=None)- Truncate results to max_lengthPrebuilt(filter_ref, input_user_id=None, name=None)- Reference to prebuilt filter
Score Configs:
Ensemble(value_model, input_user_id=None, ...)Passthrough(name=None)
Reorder Steps:
Diversity(diversity_attributes=None, strength=0.5, ...)Boosted(retriever, strength=0.5, name=None)Exploration(retriever, strength=0.5, name=None)
Execute Saved Query
result = client.execute_saved_query(
'my_engine',
'recommendations',
parameters={'userId': '123', 'limit': 20}
)
TypeScript/Node.js SDK (V2)
Installation
NPM Installation
npm install @shaped.ai/client
V2 Query API with Fluent Builders
The V2 SDK provides a fluent, type-safe query builder API:
import { Client, RankQueryBuilder } from '@shaped.ai/client';
const client = new Client({ apiKey: 'your-api-key' });
// Build a query using the fluent builder
const query = new RankQueryBuilder()
.from('item')
.retrieve(step => step
.columnOrder([{name: 'popularity', ascending: false}], {limit: 1000}))
.retrieve(step => step
.textSearch('laptop', {type: 'vector', textEmbeddingRef: 'text_emb'}))
.filter(step => step.expression('price < 1000'))
.score(step => step.ensemble('lightgbm', {inputUserId: '$parameters.userId'}))
.reorder(step => step.diversity({attribute: 'category'}))
.limit(50)
.columns(['item_id', 'title', 'price'])
.build();
// Execute the query
const result = await client.executeQuery(
'my_engine',
query,
{userId: '123'}
);
Execute Saved Query
const result = await client.executeSavedQuery(
'my_engine',
'recommendations',
{userId: '123', limit: 20}
);
Metadata
Release files for shaped 2.0.82
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| shaped-2.0.82.tar.gz | 226.5 kB | Without provenance | Details for shaped-2.0.82.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| shaped-2.0.82-py3-none-any.whl | Python 3 | none | any | Details for shaped-2.0.82-py3-none-any.whl |
Total release size: 933.2 kB
Release files / shaped-2.0.82.tar.gz
| Download URL | shaped-2.0.82.tar.gz |
|---|---|
| Size | 226.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
2f8f6e9b6eee3b9089405c495ad883f55f7f55326ae067147d450196953c32ef
|
|
BLAKE2b-256 checksum How to use checksums |
99ddfa632fd77e14d965df1a3ee9f065f7cdb188f3942ca60518fad17a36723c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Release files / shaped-2.0.82-py3-none-any.whl
| Download URL | shaped-2.0.82-py3-none-any.whl |
|---|---|
| Size | 706.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
e5b742fbbe9999f5cd2058dd558483fdea4469faefcb10340852388b1d40c549
|
|
BLAKE2b-256 checksum How to use checksums |
22cbe9281b3dc45e9620a4e4bd7d85f297aaa449b03cb7d4c0722ad1c4fb72cc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Security for shaped 2.0.82
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.