Palantir Foundry

https://www.palantir.com/docs/foundry/
Admin
Palantir Foundry is a comprehensive data integration, analytics, and application development...
Tokens:4,235,817
Snippets:34,137
Trust Score:9.7
Update:1 week ago
Show doc for...
Tokens:
Raw
### Install dependencies and start pro-code agent locally

Source: https://www.palantir.com/docs/foundry/agents/create-agent

Runs npm install and starts the agent with JSON arguments, including an example of additionalAgentContext.

```bash
npm install
npm run start -- --json-args '{ "additionalAgentContext": { "type": "string", "string": "some context" } }'
```

--------------------------------

### Start local development server

Source: https://www.palantir.com/docs/foundry/developer-console/how-to-bootstrapping-typescript

After generating your project, navigate into the project directory, install dependencies, and start the development server.

```bash
cd <project-directory>
npm install
npm run dev
```

--------------------------------

### Test live deployment via Query tab

Source: https://www.palantir.com/docs/foundry/model-integration/tutorial-productionize

Paste this example into the Query tab of the live deployment and select Run to test the model. The deployment must be initialized before querying.

```text
Copied!

```

--------------------------------

### Create and append to Iceberg table with PySpark

Source: https://www.palantir.com/docs/foundry/iceberg/jupyter

PySpark example that creates an Iceberg table with a schema and appends rows. Uses backticks to escape whitespace in the table path. Follows the Iceberg quickstart guide.

```python
from pyspark.sql.types import DoubleType, FloatType, LongType, StructType, StructField, StringType
schema = StructType([
  StructField("vendor_id", LongType(), True),
  StructField("trip_id", LongType(), True),
  StructField("trip_distance", FloatType(), True),
  StructField("fare_amount", DoubleType(), True),
  StructField("store_and_fwd_flag", StringType(), True)
])

df = spark.createDataFrame([], schema)
df.writeTo("`/.../taxis`").create()

schema = spark.table("`/.../taxis`").schema
data = [
    (1, 1000371, 1.8, 15.32, "N"),
    (2, 1000372, 2.5, 22.15, "N"),
    (2, 1000373, 0.9, 9.01, "N"),
    (1, 1000374, 8.4, 42.13, "Y")
  ]
df = spark.createDataFrame(data, schema)
df.writeTo("`/.../taxis`").append()
```

--------------------------------

### Create and insert into Iceberg table with Spark SQL

Source: https://www.palantir.com/docs/foundry/iceberg/jupyter

Spark SQL example that creates an Iceberg table partitioned by vendor_id and inserts rows. Uses backticks to escape whitespace in the table path. Follows the Iceberg quickstart guide.

```sql
spark.sql("""
        CREATE TABLE `/.../taxis`
        (
        vendor_id bigint,
        trip_id bigint,
        trip_distance float,
        fare_amount double,
        store_and_fwd_flag string
        )
        PARTITIONED BY (vendor_id);
    """)

spark.sql("""
        INSERT INTO `/.../taxis`
        VALUES (1, 1000371, 1.8, 15.32, 'N'), (2, 1000372, 2.5, 22.15, 'N'), (2, 1000373, 0.9, 9.01, 'N'), (1, 1000374, 8.4, 42.13, 'Y');
    """)
```

--------------------------------

### CREATE TABLE examples

Source: https://www.palantir.com/docs/foundry/sql-warehousing/sql-dialect

Examples of CREATE TABLE statements, including creating from a query, specifying table format, defining columns, using PARTITIONED BY, setting TBLPROPERTIES, and creating with partitioning and select.

```sql
-- examples
CREATE TABLE `/path/to/table` AS
SELECT *
FROM `/path/to/table`;

CREATE TABLE `/path/to/table` USING parquet AS
SELECT *
FROM `/path/to/table`;

CREATE TABLE `/path/to/table`
(
    id   INT,
    name STRING
);

CREATE OR REPLACE TABLE `/path/to/table` (id INT, name STRING);

CREATE TABLE `/path/to/table` (id INT, ts TIMESTAMP)
    PARTITIONED BY (days(ts));

CREATE TABLE `/path/to/table` (id INT)
    TBLPROPERTIES ('write.format.default'='parquet');

CREATE TABLE `/path/to/table`
    PARTITIONED BY (id)
    AS SELECT * FROM `/path/to/source`;
```

--------------------------------

### Get Attachment - Java

Source: https://www.palantir.com/docs/foundry/api/v2/ontologies-v2-resources/attachments/get-attachment

Java example for retrieving attachment metadata using the Get Attachment endpoint.

```java
import java.net.http.HttpClient;
import java.net.http.HttpRequest;
import java.net.http.HttpResponse;

public class GetAttachment {
    public static void main(String[] args) throws Exception {
        HttpClient client = HttpClient.newHttpClient();
        HttpRequest request = HttpRequest.newBuilder()
            .uri(URI.create("https://$HOSTNAME/api/v2/ontologies/attachments/{attachmentRid}"))
            .header(
```

--------------------------------

### Get Media Item Info - Response Example

Source: https://www.palantir.com/docs/foundry/api/v2/media-sets-v2-resources/media-sets/get-media-item-info

Example response body for the Get Media Item Info endpoint. Shows the view RID, logical timestamp, path, attribution details, and original MIME type.

```json
{
 "viewRid": "ri.mio.main.view.1",
 "logicalTimestamp": 12345,
 "path": "example.png",
 "attribution": {
   "creatorId": 1,
   "creationTimestamp": "2020-07-10 15:00:00.000"
 },
 "originallyUploadedFileMimeType": "image/png"
}
```

--------------------------------

### Set up your token

Source: https://www.palantir.com/docs/foundry/developer-console/how-to-bootstrapping-typescript

Export your token in your local environment. Below is an example using a sample personal access token, but you can generate a longer-lived one in the Developer Console. This token should not be checked into source control.

```bash
export FOUNDRY_TOKEN="sample_personal_access_token"
```

--------------------------------

### Start local development server

Source: https://www.palantir.com/docs/foundry/custom-widgets/development

Starts the local development server for custom widgets.

```bash
npm run dev
```

--------------------------------

### Get Version - Response Body

Source: https://www.palantir.com/docs/foundry/api/v2/third-party-applications-v2-resources/versions/get-version

Example response body for the Get Version endpoint. Contains the semantic version of the website.

```json
{
  "version": "1.2.0"
}
```

--------------------------------

### Example output from @osdk/cli deploy

Source: https://www.palantir.com/docs/foundry/custom-widgets/publish

Example output when publishing a production build with version 1.0.0 in the manifest. Shows the CLI version, version detection, zipping and publishing progress, and completion.

```bash
ℹ Palantir OSDK CLI 0.26.3

ℹ Found version from manifest: 1.0.0
◐ Zipping widget set files
◐ Publishing widget set files
ℹ Zipped 1.04 MiB total over 11 files
✔ Publish complete

```

--------------------------------

### GET job

Source: https://www.palantir.com/docs/foundry/compute-modules/advanced-custom-client

Documentation for the GET job endpoint, including variables, expected status codes, request example, and response parameters.

```APIDOC
## [GET] job

### Description
Retrieves a job to be processed from the internal compute module service. Returns a job payload when one exists, or a 204 status when no jobs are available.

### Method
GET

### Endpoint
$GET_JOB_URI

### Parameters
#### Headers
- **Module-Auth-Token** (string) - Required - Authentication token for the compute module service.

#### Query Parameters
- **DEFAULT_CA_PATH** (string) - Required - Path to the CA certificate used for TLS verification.

### Request Example
```
curl --header "Module-Auth-Token: $MODULE_AUTH_TOKEN" \
    --cacert $DEFAULT_CA_PATH \
    --request GET \
    $GET_JOB_URI
```

### Response
#### Success Response (200)
- **jobId** (string) - The unique identifier for the given job to be processed.
- **queryType** (string) - The name of the function to be invoked.
- **query** (object) - The payload to be passed to the function.
- **temporaryCredentialsAuthToken** (string) - A temporary token that is used with the Foundry data sidecar.
- **authHeader** (string) - The Foundry authorization token that can be used to call other services within Foundry. Only available in certain modes.

#### Response Example
```
{
    "computeModuleJobV1": {
        "jobId": "9a2a1e94-41d3-47d7-807f-db2f4c547b9c",
        "queryType": "multiply",
        "query": {
            "n": 4.0
        },
        "temporaryCredentialsAuthToken": "token-data",
        "authHeader": "auth-header"
    }
}
```

#### Status Codes
- **200:** A new job to be processed exists.
- **204:** No new jobs to be processed exist.
```

--------------------------------

### Installation

Source: https://www.palantir.com/docs/foundry/compute-modules/typescript-sdk

Install the SDK using npm or yarn.

```APIDOC
## Installation

### npm
```bash
npm install @palantir/compute-module
```

### yarn
```bash
yarn add @palantir/compute-module
```
```

--------------------------------

### foundry --help

Source: https://www.palantir.com/docs/foundry/superrepo/foundry-cli

Lists the full capabilities and options of the Foundry CLI.

```bash
foundry --help
```

--------------------------------

### Install dependencies for embeddings example

Source: https://www.palantir.com/docs/foundry/code-workspaces/palantir-provided-models

Install the required packages: palantir_models (>=0.1795.0), matplotlib, numpy, and scikit-learn. Run this in a Jupyter notebook cell.

```bash
!mamba install -y "palantir_models>=0.1795.0" matplotlib numpy scikit-learn
```

--------------------------------

### Install pandas via Task Runner

Source: https://www.palantir.com/docs/foundry/transforms-python/use-python-libraries

Manually trigger a package installation in the Task Runner by specifying the command. This example installs pandas. Enter this command in the Task Runner tab in the bottom panel.

```bash
install --packageSpecs=pandas
```

--------------------------------

### Install the latest version of your SDK with pip

Source: https://www.palantir.com/docs/foundry/developer-console/how-to-bootstrapping-python

Run this command to install the latest version of the SDK. Replace <YOUR-PACKAGE-NAME> and <INDEX-URL> with your application-specific values from the Overview page. The --extra-index-url uses the FOUNDRY_TOKEN environment variable.

```bash
pip install <YOUR-PACKAGE-NAME> --upgrade --extra-index-url "https://:$FOUNDRY_TOKEN@<INDEX-URL>"
```

--------------------------------

### SHOW FUNCTIONS examples

Source: https://www.palantir.com/docs/foundry/sql-warehousing/sql-dialect

Examples of using SHOW FUNCTIONS to list supported functions. An optional LIKE pattern filters by function name, where % matches any sequence and _ matches a single character.

```sql
SHOW FUNCTIONS;
SHOW FUNCTIONS LIKE 'abs%';
```

--------------------------------

### TypeScript v1 Function Example

Source: https://www.palantir.com/docs/foundry/getting-started/foundry-platform-summary-llm

TypeScript v1 function example. The code is incomplete in the source, showing only the imports and the start of a class with a @Function() decorator.

```typescript
import { Function } from "@foundry/functions-api";
import { Objects } from "@foundry/ontology-api";

export class FlightFunctions {
    @Function()
```
Tokens:
Raw