Prometheux

https://docs.prometheux.ai/
Admin
Prometheux is an ontology-native data processing engine that allows you to define executable...
Tokens:157,573
Snippets:1,878
Trust Score:6
Update:4 months ago
Show doc for...
Tokens:
Raw
### Cloud-Specific Examples

Source: https://docs.prometheux.ai/getting-started/installation/databricks/databricks-integration

Region examples for specific cloud providers.

```json
{"region": "us-east-1"}
```

```json
{"region": "eastus"}
```

```json
{"region": "us-central1"}
```

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

### Manual Installation - pipx

Source: https://docs.prometheux.ai/api

Install prometheux-mcp using pipx for manual setup.

```bash
pipx install prometheux-mcp
```

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

### Graph Analytics Example in Databricks Notebook

Source: https://docs.prometheux.ai/getting-started/installation/databricks/databricks-integration

Compute all shortest paths between airports using flight route data in a Databricks notebook.

```Scala
%scala  
val tableName = "air_routes_edges"  
val outputTable = "all_shortest_paths_routes"  
val program = s"""  
  @bind("air_routes","databricks inCluster=true","","prometheux_workspace.default.$tableName").  
  edge(From, To, Dist) <-   
    air_routes(Id, From, To,Label,Dist).  
  all_shortest_paths_routes(From,To,Distance) <- #ASP(edge).  
  @output("all_shortest_paths_routes").   
  @bind("all_shortest_paths_routes","databricks inCluster=true","","$outputTable").  
"""  
val args = Array(program)  
uk.co.prometheux.prometheuxreasoner.PrometheuxReasonerMain.main(args)  
```

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

### Get Usage Status cURL Example

Source: https://docs.prometheux.ai/api/config

Example of how to retrieve API usage statistics and limits using a cURL command.

```bash
curl -X GET "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/users/usage-status" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN"
```

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

### Python SDK Installation

Source: https://docs.prometheux.ai/api

Install or upgrade the Prometheux Chain Python SDK.

```bash
pip install --upgrade prometheux_chain
```

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

### SQL Queries with Ordering and Distinct - Example: ORDER BY date

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example: ORDER BY date.

```vadalog
% Example: ORDER BY date  

@bind("orders", "json", "data/orders", "orders.json").  

  

recent_orders() <- SELECT order_id, customer.name AS customer_name, order_date, total_amount   

                   FROM orders   

                   ORDER BY order_date DESC.  

  

@output("recent_orders").  

```

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

### Install into Jupyter - Install Package

Source: https://docs.prometheux.ai/api

Install the prometheux_chain package in editable mode after cloning.

```bash
!cd prometheux_chain && pip install -e .
```

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

### Windows Installation Script

Source: https://docs.prometheux.ai/api/mcp/local

Automated installation script for Windows.

```powershell
# Windows
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/prometheuxresearch/px-mcp-server/main/install.ps1" -OutFile "install.ps1"; .\install.ps1
```

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

### Explanation Output Structure Example

Source: https://docs.prometheux.ai/learn/vadalog/explanations

This example demonstrates the VadaLog code for deriving facts and enabling explanations to the console.

```VadaLog
edge(1, 2).

edge(2, 3).

edge(3, 4).

tc(X, Y) <- edge(X, Y).

tc(X, Z) <- tc(X, Y), edge(Y, Z).

@output("tc").

@explain("console").
```

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

### Example 1 Output

Source: https://docs.prometheux.ai/learn/vadalog/sql-integration

The expected output for Example 1, showing the 'adults' facts.

```text
adults("Alice", 25)  
adults("Bob", 30)  

```

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

### Install into Jupyter - Clone Repo

Source: https://docs.prometheux.ai/api

Clone the prometheux_chain repository for manual installation in Jupyter.

```bash
!git clone https://github.com/prometheuxresearch/prometheux_chain.git
```

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

### Get Bitcoin Price

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Quick start example to get the current Bitcoin price from the CoinGecko API.

```Vadalog
@bind("btc", "api", "https://api.coingecko.com/api/v3/", "coins/bitcoin").  
result() <- SELECT id, symbol, market_data.current_price.usd AS price FROM btc.  
@output("result").  


```

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

### Complete Configuration Example

Source: https://docs.prometheux.ai/api/config

A comprehensive example demonstrating how to set up authentication, configure the backend URL, and set LLM parameters using the Prometheux AI SDK.

```python
import prometheux_chain as px
import os

# Set up authentication via environment variable
os.environ['PMTX_TOKEN'] = 'my_pmtx_token'

# Or set via SDK
px.config.set("PMTX_TOKEN", "my_pmtx_token")

# Configure the backend URL
px.config.set('JARVISPY_URL', "https://api.prometheux.ai/jarvispy/my-org/my-user")

# Configure LLM settings
px.config.update_config({
    "LLM_PROVIDER": "OpenAI",
    "LLM_API_KEY": "your_openai_api_key",
    "LLM_VERSION": "gpt-4o",
    "LLM_TEMPERATURE": 0.7,
    "LLM_MAX_TOKENS": 4000
})

# Verify configuration
print(f"API URL: {px.config.get('JARVISPY_URL')}")
print(f"LLM Provider: {px.config.get('LLM_PROVIDER')}")
```

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

### First API Call Example

Source: https://docs.prometheux.ai/api

Example of making a POST request to save a project.

```curl
curl -X POST "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/projects/save" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -d '{
    "project": {
      "name": "my_first_project"
    }
  }'
```

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

### Save User Config cURL Example

Source: https://docs.prometheux.ai/api/config

Example of how to save user configuration using a cURL command.

```bash
curl -X POST "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/users/save-config" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -d '{ 
    "config_data": {
      "theme": "dark"
    },
    "scope": "user"
  }'
```

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

### Get Football Team Data

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Quick start example to get football team data from TheSportsDB API.

```Vadalog
@bind("teams", "api", "https://www.thesportsdb.com/api/v1/json/3/", "searchteams.php?t=Arsenal").  
result() <- SELECT SIZE(teams) AS team_count FROM teams.  
@output("result").  


```

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

### Advanced Configuration with Sort Key and GSI

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates creating a table with both partition and sort keys, plus a Global Secondary Index.

```vadalog
% Bind to CSV file containing order data  

@bind("orders_csv", "csv useHeaders=true", "disk/data/input", "orders.csv").  

  

% Define the data model for orders  

@model("orders_dynamodb", "['customer_id:string', 'order_date:string', 'order_id:string', 'total_amount:double', 'status:string']").  

  

% Define a rule to map CSV data to the DynamoDB concept  

orders_dynamodb(CustomerId, OrderDate, OrderId, TotalAmount, Status) <-
        orders_csv(CustomerId, OrderDate, OrderId, TotalAmount, Status).  

  

% Declare the output concept  

@output("orders_dynamodb").  

  

% Bind with advanced configuration: sort key and Global Secondary Index  

@bind("orders_dynamodb",   
      "dynamodb region='us-east-1', username='AKIAIOSFODNN7EXAMPLE', password='wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY', partitionKey='customer_id', sortKey='order_date', billingMode='PROVISIONED', readCapacity=10, writeCapacity=5, secondaryIndexName='StatusIndex', secondaryIndexPartitionKey='status'",   
      "", "orders").  


```

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

### Get Engine Status Response Example

Source: https://docs.prometheux.ai/api/endpoints/vadalog

Example JSON response for the Get Engine Status endpoint.

```json
{
  "data": {
    "status": "ready",
    "timestamp": "2026-03-30T12:00:00.000000"
  },
  "message": "Evaluation service status retrieved successfully",
  "status": "success"
}
```

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

### Import Workspace cURL Example

Source: https://docs.prometheux.ai/api/endpoints/projects

Example of how to import workspace data using the cURL command.

```shell
curl -X POST "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/projects/import-workspace" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN" \
  -d '{  
    "export_data": {...},  
    "scope": "user"  
  }'  
```

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

### Using ask() with Inline Connection Options

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

An example of overriding Qdrant connection settings inline with the ask() function.

```text
ask("search query", "host=52.8.114.148,port=6334,apiKey=my-key,collection=my_collection")
```

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

### macOS/Linux Installation Script

Source: https://docs.prometheux.ai/api/mcp/local

Automated installation script for macOS and Linux.

```bash
# macOS/Linux
curl -sSL https://raw.githubusercontent.com/prometheuxresearch/px-mcp-server/main/install.sh -o install.sh && chmod +x install.sh && ./install.sh
```

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

### Minimal Qdrant Read Configuration

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates the simplest @bind for a semantic search read from Qdrant, utilizing default platform configurations.

```VadaLog
@bind("my_predicate", "qdrant query='search text'", "", "my_collection").
```

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

### Create a JSON key file for Google Cloud

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Command to create a service account key file for Google Cloud authentication.

```bash
gcloud iam service-accounts keys create ~/gcp-credentials.json   --iam-account="$SA_NAME@$PROJECT_ID.iam.gserviceaccount.com"
```

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

### Get Engine Status

Source: https://docs.prometheux.ai/api/endpoints/vadalog

Example using curl to get the status of the Vadalog evaluation service.

```curl
curl -X GET "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/vadalog/status" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN"
```

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

### Get User Role cURL Example

Source: https://docs.prometheux.ai/api/config

Example of how to retrieve a user's role using a cURL command.

```bash
curl -X GET "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/users/get-role" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN"
```

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

### Vadalog Examples - Full

Source: https://docs.prometheux.ai/learn/vadalog/annotations

A comprehensive example demonstrating state, event, and out_event models, including data instantiation and a rule.

```vadalog
@model("state","['state_id(ID):string', 'Type:string', 'Balance:double']").  
@model("event","['Id:int', 'StartState:state', 'EndState:state', 'Prob:double']").  
% The next line is optional, since we're not using it an any computation.  
@model("out_event","['Start State(ID):string', 'Type:string', 'Balance:double']").   
  
event(1, "start state A", "end state A", 0.1).  
state("start state A", "positive", 10.0).  
state("end state A", "negative", -10.0).  
  
out_event(StartState, EndState, Prob) <- event(Id, StartState, EndState, Prob).  
@output("out_event").  

```

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

### Populating PostgreSQL from CSV

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows how to read data from a CSV file and populate a customer table in a PostgreSQL database.

```vadalog
% Bind the 'customer_postgres_csv' concept to the CSV file located in 'disk/data/generated_data/customer_postgres.csv'  

% The option 'useHeaders=true' indicates the CSV file contains headers  

@bind("customer_postgres_csv", "csv useHeaders=true", "disk/data/generated_data", "customer_postgres.csv").  

  

% Define a rule that extracts CustomerID, Name, Surname, and Email from the CSV and assigns them to the 'customer_postgres' concept  

customer_postgres(CustomerID, Name, Surname, Email) <-
        customer_postgres_csv(CustomerID, Name, Surname, Email).  

  

% Define the data model for the 'customer_postgres' concept (mapping column names to types)  

@model("customer_postgres", "['customer_id:int', 'name:string', 'surname:string', 'email:string']").  

  

% Declare the 'customer_postgres' concept as the output, which will be written to the PostgreSQL database  

@output("customer_postgres").  

  

% Bind the 'customer_postgres' concept to a PostgreSQL table 'customer' in the 'prometheux' database  

% Specify the database connection details (host, port, username, and password)  

@bind("customer_postgres", "postgresql host='postgres-host', port=5432, username='prometheux', password='myPassw'",   

      "prometheux", "customer").  

  

```

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

### struct:get() Example

Source: https://docs.prometheux.ai/learn/vadalog/expressions/specialized-functions

Demonstrates extracting a field value from a struct by its field name using struct:get().

```Vadalog
a(10, 20).

b(S) <- a(X, Y), S = struct("field1", X, "field2", Y).

c(F1, F2) <- b(S), F1 = struct:get("field1", S), F2 = struct:get("field2", S).

@output("c").
```

```Vadalog
c(10, 20).
```

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

### Sample Call for /config-info/setAll

Source: https://docs.prometheux.ai/api/rest-api

Example cURL command to set multiple key-value configuration pairs.

```bash
curl -X POST "http://localhost:8080/config-info/setAll" \
  -H "Content-Type: application/json" \
  -d '{
    "database.password": "myDBPassword",
    "s3aAccessKey": "myAWSAccessKey",
    "s3aSecretKey": "myAWSSecretKey"
  }'
```

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

### Extracting multiple results using collections:get and struct:get

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows how to retrieve the top 3 similar cities from a collection and extract specific fields (name) from each of the results using `collections:get` and `struct:get`.

```VadaLog
% Get the top 3 similar cities and extract each one  
city_query("European capital").  
  
  
top3(Query, Name1, Name2, Name3) <- city_query(Query),  
    Answer = ask(Query, "collection=cities,limit=3"),  

    L = as_list(Answer, "name:string, __score__:double"),  

    First = collections:get(L, 1),  

    Second = collections:get(L, 2),  

    Third = collections:get(L, 3),  

    Name1 = struct:get("name", First),  

    Name2 = struct:get("name", Second),  

    Name3 = struct:get("name", Third).  

@output("top3").  


```

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

### Complete Workflow Python SDK Example

Source: https://docs.prometheux.ai/api/endpoints/projects

A comprehensive example demonstrating the workflow of creating, listing, loading, and cleaning up projects using the Python SDK.

```python
import prometheux_chain as px  
import os  
  
# Set up authentication and configuration  
os.environ['PMTX_TOKEN'] = 'my_pmtx_token'  
px.config.set('JARVISPY_URL', "https://api.prometheux.ai/jarvispy/my-org/my-user")  
  
# Create a new project  
project_id = px.save_project(project_name="test_project")  
  
# List all projects  
projects = px.list_projects()  
print(f"Available projects: {projects}")  
  
# Load the created project  
project_data = px.load_project(project_id)  
  
# Clean up when done  
px.cleanup_projects(project_id=project_id)  
```

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

### Semantic Search with ASK

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Equivalent example using the ASK syntax, demonstrating its interchangeability with SELECT for semantic search.

```VadaLog
result(Id, Category, Name, Price, Score) <-   
    ASK "wireless noise cancelling headphones" FROM products LIMIT 5.  

```

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

### Filtering with struct:get

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates how to filter orders by status using the struct:get function to access nested fields.

```vadalog
@bind("orders", "json", "data/orders", "orders.json").  

% Filter orders by status using struct:get  
shipped_orders(OrderId, CustomerName) <-
    orders(OrderId, _, Customer, _, _, _, Status),   
    OrderId = struct:get("order_id", orders),  
    CustomerName = struct:get("name", Customer),   
    Status = "shipped".  

@output("shipped_orders").  

```

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

### Combining preview and count

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example demonstrating a common pattern of running both a paginated preview and a count in the same program.

```VadaLog
@bind("order_rel", "neo4j username='neo4j', password='myPassw', host='neo4j-host', port=7680", "neo4j", "(:Order)-[IS_RELATED_TO]->(:Person)").  

  

% Preview: first 10 results  

order_preview() <- SELECT * FROM order_rel LIMIT 10 OFFSET 0.  

  

% Count: total number of relationships  

order_count(Total) <- SELECT COUNT(*) AS total FROM order_rel.  

  

@output("order_preview").  

@output("order_count").  


```

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

### Accessing Nested Fields with struct:get

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example of using the `struct:get` function in Vadalog to access nested fields within a JSON structure, specifically binding and extracting a 'MetricName' from a 'MetricStruct'.

```Vadalog
% Access nested fields after binding  

result(MetricName) <-   

    api_data(..., MetricStruct, ...),  

    MetricName = struct:get("instance", MetricStruct).  

```

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

### Using ask() with Collection Option

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

An example of using the ask() function with the 'collection' option, falling back to pmtx.properties for other settings.

```text
ask("search query", "collection=my_collection,limit=5")
```

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

### Get Configuration Value - cURL Example

Source: https://docs.prometheux.ai/api/config

Retrieves a value from the configuration using the specified key via REST API.

```bash
curl -X GET "https://api.prometheux.ai/jarvispy/my-org/my-user/api/v1/users/load-config?scope=user" \
  -H "Authorization: Bearer YOUR_JWT_TOKEN"
```

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

### Load User Config Python SDK Example

Source: https://docs.prometheux.ai/api/config

Example of how to load user-scoped configuration using the Python SDK.

```python
import prometheux_chain as px

config = px.load_user_config(scope="user")
```

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

### Query Prometheus Metrics

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Quick start example to query Prometheus metrics using the API.

```Vadalog
@bind("metrics", "api", "http://prometheus-server:9090/api/v1/", "query?query=up").  
result(Status, ResultType) <- metrics(Data, Status),   
                              ResultType = struct:get("resultType", Data).  
@output("result").  


```

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

### Composition Example - Event

Source: https://docs.prometheux.ai/learn/vadalog/annotations

Defines an 'event' predicate that uses the 'state' predicate for 'Start State' and 'End State' fields.

```vadalog
@model("event", "['Id:int', 'Start State:state', 'End State:state', 'Prob:double']").  

```

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

### Install into Jupyter - Copy to Site-packages

Source: https://docs.prometheux.ai/api

Copy the prometheux_chain project into site-packages for importability.

```bash
!cd prometheux_chain && cp -r prometheux_chain /opt/app-root/lib/python3.x/site-packages
```

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

### Combining Both in a Single Program

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Demonstrates mixing SELECT * queries with aggregate queries over the same Qdrant predicate in a single program.

```vadalog
@bind("mol_in", "qdrant", "", "molecules").  

  
% Read a page of results  
mol_page() <- SELECT * FROM mol_in LIMIT 10 OFFSET 0.  
@output("mol_page").  

  
% Count all rows  
mol_count(Total) <- SELECT COUNT(*) as total FROM mol_in.  
@output("mol_count").  

  
% Average weight  
mol_avg(AvgWeight) <- SELECT AVG(weight) as avg_weight FROM mol_in.  
@output("mol_avg").  

```

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

### Neo4j Docker Setup with APOC

Source: https://docs.prometheux.ai/learn/vadalog/chase-provenance

This Docker script sets up a Neo4j container instance with the APOC library installed.

```bash
#!/bin/bash  

docker run --rm -d --name neo4j-vada \
-p 7474:7474 \
-p 7687:7687 \
--volume=${PWD}/neo4j-data:/var/lib/neo4j/data \
--volume=${PWD}/neo4j-import:/var/lib/neo4j/import \
--volume=${PWD}/neo4j-plugins:/var/lib/neo4j/plugins \
-e NEO4J_apoc_export_file_enabled=true \
-e NEO4J_apoc_import_file_enabled=true \
-e NEO4J_apoc_import_file_use__neo4j__config=true \
-e NEO4 জৈLABS_PLUGINS='["graph-data-science","apoc"]' \
-e NEO4J_dbms_memory_heap_max__size=2G \
-e NEO4J_dbms_security_procedures_unrestricted=gds.\* neo4j:4.4.31  

```

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

### `substring` operator example

Source: https://docs.prometheux.ai/learn/vadalog/expressions/operators

A rule using `substring` returns a substring from the specified `start` to `end` index, using zero-based indexing.

```Vadalog
q(K1, K2, Kn, J) <- body, J = substring(x, start, end).  

```

```Vadalog
a("prometheux").  

b("oxford").  

q(Y,J) <- a(X), b(Y), J = substring(X,4,10).  
@output("q").  

```

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

### Example 1: Simple API Call (JSON - Default Format)

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates a simple API call to fetch user data from JSONPlaceholder, utilizing the default JSON response format and querying nested fields with SQL.

```vadalog
% Bind to a public API that returns JSON (no authentication needed)  

% responseFormat defaults to JSON  

@bind("users_api", "api", "https://jsonplaceholder.typicode.com/", "users").  

  

% Use SQL in rule body to query nested fields  

result() <- SELECT name, email, address.city AS city   
            FROM users_api   
            WHERE id <= 5.  
  

@output("result").  

```

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

### End-to-End Pipeline — Write, Query, and Reason (Step 1)

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This is the first step of an end-to-end pipeline, demonstrating writing product data with concept-aware embeddings to Qdrant.

```Vadalog
% Step 1: Write product data with concept-aware embeddings  

product("Wireless Headphones", "Electronics", 79.99).
product("Desk Lamp LED", "Lighting", 34.99).
product("Mechanical Keyboard", "Electronics", 129.99).
product("Standing Desk", "Furniture", 499.99).
  

product_out(Name, Category, Price) <-
    product(Name, Category, Price).
@output("product_out").  

  

@bind("product_out",   

      "qdrant host='localhost', port=6334, vectorDimension=3072",   

      "", "products").  

@model("product_out",   

       "['name:string', 'category:string', 'price:double']",  

       "[name] is a [category] product priced at [price] dollars").  



```

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

### Simple Example: Reading JSON Files

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates reading a JSON file without any query, binding the 'orders' concept to a JSON file and then accessing the data in Vadalog rules.

```vadalog
% Bind the 'orders' concept to a JSON file  

@bind("orders", "json", "data/orders", "orders.json").  

  

% Access the data in Vadalog rules  

all_orders(OrderId, CustomerName, Total) <- orders(OrderId, _, CustomerName, _, _, Total, _).  

  

@output("all_orders").  

```

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

### Keeping nested groups as structs

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows how to keep nested COBOL groups as structs, allowing them to be queried with `struct:get`.

```VadaLog
@bind("customer_master",  

      "cobol copybook='/data/cobol/customer_master.cpy',  

             cobolPreset='flat-fixed-ebcdic',  

             cobolFlattenPolicy='keepNested'",  

      "/data/cobol",  

      "customer_master.dat").  

  

% CUST-ADDRESS is now a struct column; reach into it with struct:get.  
customer_city(Id, City) <-  

    customer_master(Id, _, _, Address, _, _, _, _),  

    City = struct:get("CUST_CITY", Address).  

  

@output("customer_city").  


```

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

### Python Example: Starting a Conversation

Source: https://docs.prometheux.ai/api/chat-api

Demonstrates how to initiate a conversation with the Chat API using Python, including handling streaming responses and adding to conversation history.

```python
conversation = [
  
    {"role": "user", "content": "How do I use LLM functions and embeddings in Vadalog?"}
  
]
  
response = requests.post(url, json={"messages": conversation}, stream=True)
  
assistant_response = ""
for chunk in response.iter_content(chunk_size=1024, decode_unicode=True):
    if chunk:
        lines = chunk.split('\n')
        for line in lines:
            if line.startswith('0:"') and line.endswith('"'):
                assistant_response += line[3:-1]
  
# Add to conversation for follow-up
conversation.append({"role": "assistant", "content": assistant_response})

print("AI Functions Example:", assistant_response)
  
# Follow-up question
conversation.append({"role": "user", "content": "Can you show me a similarity analysis example?"})
follow_up = requests.post(url, json={"messages": conversation}, stream=True)
# Process follow-up response...
```

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

### Reading from Text File

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example demonstrating how to read from a text file, define a data model for a 'location' concept, bind it to a text file, and extract Name and Description.

```Vadalog
% Define the data model for the 'location' concept  

@model("location","['Name:string', 'Description:string']").  

  

% Bind the 'location' concept to the text file  

@bind("location","text","/path/to/file","hansel_gretel_excerpt.txt").  

  

% Define a rule to extract Name and Description from the text file  

location_head(Name,Description) <- location(Name,Description).  

  

% Declare the output concept 'location_head' for making the processed data available  

@output("location_head").  

```

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

### Using `collections:explode` to Convert Arrays to Rows

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates how to use `collections:explode` to convert an array of league objects into individual rows, and then extract specific fields using `struct:get`.

```vadalog
% Get all football leagues from TheSportsDB  

@bind("leagues_api", "api", "https://www.thesportsdb.com/api/v1/json/3/", "all_leagues.php").  
  

% Explode the leagues array into individual rows  
leagues_linear(League) <- leagues_api(LeaguesArray),   
                          League = collections:explode(LeaguesArray).  
  

% Extract fields from each league struct using struct:get  
result(LeagueId, LeagueName) <- leagues_linear(League),   
                                 LeagueId = struct:get("idLeague", League),   
                                 LeagueName = struct:get("strLeague", League).  
  

@output("result").  

```

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

### Minimal Qdrant Write Configuration

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows the simplest @bind for a Qdrant write operation with concept-aware embeddings, relying on default platform configurations.

```VadaLog
@bind("my_predicate", "qdrant", "", "my_collection").
```

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

### CSV Bind Command Example with Configuration

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

An example of the @bind command for CSV with specific configuration options like delimiter and record separator.

```vadalog
@bind("relation",  
      "csv useHeaders=true, delimiter='\t', recordSeparator='\n'",  
      "filepath",  
      "filename").  

```

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

### Using `collections:transform` with Lambda Expressions

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows how to use `collections:transform` with a lambda expression to transform each league struct into a simpler array `[id, name]` before exploding it into rows. It then accesses array elements using `collections:get`.

```vadalog
% Get all football leagues  

@bind("leagues_api", "api", "https://www.thesportsdb.com/api/v1/json/3/", "all_leagues.php").  
  

% Transform each league struct into a simpler array [id, name]  
leagues_transformed(TransformedArray) <-   
    leagues_api(LeaguesArray),   
    TransformedArray = collections:transform(LeaguesArray, "x -> array(x.idLeague, x.strLeague)").  


% Explode the transformed array into rows  
leagues_linear(LeagueData) <-   
    leagues_transformed(TransformedArray),   
    LeagueData = collections:explode(TransformedArray).  


% Access array elements using collections:get (1-indexed)  
result(LeagueId, LeagueName) <-   
    leagues_linear(LeagueData),   
    LeagueId = collections:get(LeagueData, 1),   
    LeagueName = collections:get(LeagueData, 2).  


@output("result").  

```

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

### Reading Full Customer Table from PostgreSQL

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates reading the full customer table from PostgreSQL.

```vadalog
% Bind the 'customer_postgres' concept to the PostgreSQL table 'customer' in the 'prometheux' database  

@bind("customer_postgres", "postgresql host='postgres-host', port=5432, username='prometheux', password='myPassw'",   

      "prometheux", "customer").  

  

% Define a rule to extract CustomerID, Name, Surname, and Email from the 'customer' table in PostgreSQL  

customer_postgres_test(CustomerID, Name, Surname, Email) <-
        customer_postgres(CustomerID, Name, Surname, Email).  

  

% Declare the output concept 'customer_postgres_test' to make the processed data available  

@output("customer_postgres_test").  

  

```

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

### Python SDK Installation

Source: https://docs.prometheux.ai/getting-started/installation/cloud-platform

Command to install the Prometheux Python SDK.

```bash
pip install prometheux_chain
```

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

### Example JSON Structure

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This is an example of an e-commerce orders JSON structure used in Prometheux examples.

```JSON
[
  {
    "order_id": "ORD-2025-001",
    "order_date": "2025-01-15",
    "customer": {
      "customer_id": "CUST-123",
      "name": "Alice Johnson",
      "email": "alice@example.com"
    },
    "shipping_address": {
      "street": "123 Main St",
      "city": "Boston",
      "state": "MA",
      "zip": "02101"
    },
    "items": [
      {"product": "Laptop", "quantity": 1, "price": 1299.99},
      {"product": "Mouse", "quantity": 2, "price": 29.99}
    ],
    "total_amount": 1359.97,
    "status": "shipped"
  }
]

```

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

### Fact Examples

Source: https://docs.prometheux.ai/learn/vadalog/grammar

Examples of fact declarations.

```vadalog
person("alice", 30).

employee("bob", "engineering", 75000).

company("acme", "technology").
```

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

### Syntax for ask() Function

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

The basic syntax for the ask() function, which performs Qdrant vector search per-row.

```text
ask(prompt_or_vector, "option1=value1,option2=value2")
```

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

### Manual Installation using pipx (Windows)

Source: https://docs.prometheux.ai/api/mcp/local

Manual installation steps for prometheux-mcp on Windows using pipx, which installs the package in an isolated environment.

```powershell
pip install pipx
pipx ensurepath
pipx install prometheux-mcp
```

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

### @bind Example

Source: https://docs.prometheux.ai/learn/vadalog/annotations

An example showing how to use @bind to specify data sources for 'm' and 'q' atoms, followed by a rule that joins them.

```vadalog
@output("m").  
@bind("m","postgres","doctors_source","Medprescriptions").  
@bind("q","sqlite","doctors_source","Medprescriptions").  
m(X) <- b(X),q(X).  

```

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

### Writing to S3

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example demonstrates how to read data from a local CSV file and write it to an S3 bucket.

```Vadalog
% Bind the 'user_csv' concept to the local CSV file 'users.csv' located in 'disk/datasources'  

@bind("user_csv", "csv useHeaders=true", "disk/datasources", "users.csv").  

  

% Define a rule to map the data from 'user_csv' to the 'user' concept  

user(X) <- user_csv(X).  

  

% Declare the output concept 'user' for writing the data to an S3 bucket  

@output("user").  

  

% Bind the 'user' concept to a CSV file in the specified S3 bucket  

@bind("user", "csv", "s3a://your-s3-bucket/", "user.csv").  

  

```

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

### Manual Installation using pipx (macOS)

Source: https://docs.prometheux.ai/api/mcp/local

Manual installation steps for prometheux-mcp on macOS using pipx, which installs the package in an isolated environment.

```bash
brew install pipx
pipx ensurepath
pipx install prometheux-mcp
```

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

### Min post-processing example 1

Source: https://docs.prometheux.ai/learn/vadalog/annotations

Demonstrates 'min' post-processing on atom 'p', calculating the minimum value for the third position, grouped by other positions.

```vadalog
t(1,"b",5).

t(1,"b",1).

t(1,"c",1).

p(X,Y,Z) <- t(X,Y,Z).

@output("p").

@post("p","min(3)").
```

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

### collections:get()

Source: https://docs.prometheux.ai/learn/vadalog/expressions/collections

Get an element at a specific index (1-based).

```vadalog
collections:get(arr: array<T>, index: int) 
→ T
```

```vadalog
first_skill(Name, FirstSkill) <-

    employee(Name, Skills),

    FirstSkill = collections:get(Skills, 1).
```

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

### Literal Examples

Source: https://docs.prometheux.ai/learn/vadalog/grammar

Examples of positive, negative, and domain literals.

```vadalog
% Positive literal
employee(Name, Dept) <- person(Name), works_in(Name, Dept).

% Negative literal
available(Name) <- employee(Name), not on_vacation(Name).

% Domain literal
all_values(X) <- dom(*), X = value.
```

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

### cURL Example

Source: https://docs.prometheux.ai/api/chat-api

Example of how to interact with the Vadalog API using cURL.

```curl
curl -X POST https://chat-docs.prometheux.ai/api/vadalog \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Show me CSV file processing",
    "context": "data integration", 
    "include_docs": true
  }'



```

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

### Reading from PDF File

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example demonstrating how to read from a PDF file, bind a 'person' concept to it, and extract Name and Role.

```Vadalog
% Bind the 'person' concept to the PDF file  

@bind("person","binaryfile","/path/to/file","hansel_gretel_excerpt.pdf").  

  

% Define a rule to extract Name and Role from the PDF file  

person_head(Name,Role) <- person(Name,Role).  

  

% Declare the output concept 'person_head' for making the processed data available  

@output("person_head").  

```

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

### SQL Queries with Filtering and Aggregation - Example 2: Aggregation with GROUP BY

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example 2: Aggregation with GROUP BY.

```vadalog
% Example 2: Aggregation with GROUP BY  

@bind("orders", "json", "data/orders", "orders.json").  

  

orders_by_city() <- SELECT shipping_address.city AS city,   

                           COUNT(*) AS order_count,  

                           SUM(total_amount) AS total_revenue  

                    FROM orders   

                    GROUP BY shipping_address.city.  

  

@output("orders_by_city").  

```

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

### Python Example

Source: https://docs.prometheux.ai/api/chat-api

Example of how to interact with the Vadalog API using Python.

```python
import requests
  
url = "https://chat-docs.prometheux.ai/api/vadalog"
payload = {
    "query": "Show me CSV file processing",
    "context": "data integration",
    "include_docs": True
}
  
response = requests.post(url, json=payload)
data = response.json()
  
print(f"Response: {data['response']}")
print(f"Code examples: {len(data['code_examples'])}")
print(f"Related docs: {len(data['relevant_docs'])}")



```

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

### Sample Call with modules for /evaluate endpoint

Source: https://docs.prometheux.ai/api/rest-api

Example of calling the /evaluate endpoint with multiple modules.

```bash
curl 'http://localhost:8080/evaluate?program=@module(%22m1%22).@include(%22m2%22).a(1).&otherPrograms=@module(%22m2%22).b(X)<-a(X).@output(%22b%22).&modules=@module(%22m3%22).b(X)<-a(X).@output(%22b%22).'
```

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

### Update Installation - Remove Existing

Source: https://docs.prometheux.ai/api

Remove an existing prometheux_chain installation from site-packages.

```bash
!rm -r /opt/app-root/lib/python3.x/site-packages/prometheux_chain
```

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

### Example usage of @model and @output

Source: https://docs.prometheux.ai/learn/vadalog/annotations

An example demonstrating the usage of @model and @output annotations.

```vadalog
b(1, "2", 1.0, "Davide").  
a(A, B, C, D) <- b(A, B, C, D).  
  
@model("a", "['first:string', 'second:string', 'third:double', 'fourth:string']").  
@output("a").  
```

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

### Input configuration for Unity Catalog tables

Source: https://docs.prometheux.ai/getting-started/installation/databricks/databricks-integration

Configure input tables using Unity Catalog three-level namespace.

```Vadalog
% Input configuration for Unity Catalog tables  
@bind("customers", "databricks", "sales_catalog.crm_schema", "customers").  
@bind("orders", "databricks", "sales_catalog.transactions_schema", "orders").  
% Vadalog rule using Unity Catalog data  
customer_orders(CustomerName, OrderId, OrderDate, Amount) <-   
    customers(CustomerId, CustomerName, Email),  
    orders(OrderId, CustomerId, OrderDate, Amount).  
@output("customer_orders").  
@bind("customer_orders", "databricks", "analytics_catalog.reports_schema", "customer_orders").  
```

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

### Using Parameters for Dynamic Time Ranges

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Demonstrates a range query using parameterized start time, end time, and step for dynamic data retrieval.

```vadalog
% Define time range parameters (epoch timestamps)  
% Example: Large time range from 2026-2028 for long-term validity  
@param("start_time", "1770700000").  % Feb 10, 2026  
@param("end_time", "1770800000").    % Feb 11, 2026    
@param("step", "60").  
  
% Range query with parameterized time range  
@bind("metric_range",   
      "api authType=bearer, token=${PROMETHEUS_TOKEN}",   
      "https://prometheus.example.com/api/v1/",   
      "query_range?query=up&start=${start_time}&end=${end_time}&step=${step}").  
  
% Process the time series data  
time_series_data() <- SELECT status,   
                             data.resultType AS result_type,  
                             SIZE(data.result) AS total_series  
                      FROM metric_range  
                      WHERE status = 'success'.  
  
@output("time_series_data").  

```

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

### Rule Head Examples

Source: https://docs.prometheux.ai/learn/vadalog/grammar

Examples of different rule head types.

```vadalog
% Regular head
adult_customer(Name) <- customer(Name, Age), Age >= 18.

% Multiple heads
high_value(Name), vip(Name) <- customer(Name, _, Balance), Balance > 10000.

% EGD head
X = Y <- person(X, Age1), person(Y, Age2), Age1 = Age2.

% Constraint (false head)
#F <- employee(Name, Dept), Dept = "invalid".
```

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

### Min post-processing example 2

Source: https://docs.prometheux.ai/learn/vadalog/annotations

Illustrates 'min' post-processing on atom 'q' with multiple positions (1 and 2), showing lexicographical comparison for minimum tuple selection.

```vadalog
t(1,"b",1).

t(2,"c",1).

t(1,"a",1).

q(X,Y,Z) <- t(X,Y,Z).

@output("q").

@post("q","min(1,2)").
```

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

### Error Response Example

Source: https://docs.prometheux.ai/api

Example of an error response body from the REST API.

```json
{
  "data": null,
  "message": "Project not found",
  "status": "error"
}
```

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

### JavaScript/TypeScript Example

Source: https://docs.prometheux.ai/api/chat-api

Example of how to interact with the Vadalog API using JavaScript or TypeScript.

```javascript
const response = await fetch('https://chat-docs.prometheux.ai/api/vadalog', {
  
  method: 'POST',
  
  headers: {
    
    'Content-Type': 'application/json',
    
  },
  
  body: JSON.stringify({
    
    query: 'How do I use aggregations in Vadalog?',
    
    include_docs: true
    
  })
  
});
  
  
const data = await response.json();
console.log(data.response);        // Text response
console.log(data.code_examples);   // Extracted code blocks
console.log(data.relevant_docs);   // Related documentation



```

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

### Example 2: Querying Nested JSON with SQL

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

This example shows how to bind to the GitHub API using bearer authentication and query nested JSON structures, including array elements, with SQL.

```vadalog
% Bind to GitHub API with authentication  

@bind("github_issues",   
      "api authType=bearer, token=${GITHUB_TOKEN}",   
      "https://api.github.com/repos/", "owner/repo/issues").  

  

% Query nested JSON structures with SQL  

open_bugs() <- SELECT title,   
                      user.login AS author,   
                      labels[0].name AS primary_label,  
                      created_at  
               FROM github_issues   
               WHERE state = 'open'   
                 AND labels[0].name = 'bug'  
               ORDER BY created_at DESC.  
  

@output("open_bugs").  

```

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

### Multi-Step Analysis in Databricks Notebook

Source: https://docs.prometheux.ai/getting-started/installation/databricks/databricks-integration

Perform multi-step analysis: data preprocessing followed by graph analysis.

```Scala
%scala  
// Step 1: Data preprocessing  
val preprocessProgram = s"""  
  @bind("raw_routes","databricks inCluster=true","","raw_flight_data").  
    
  clean_routes(From, To, Distance) <-   
    raw_routes(_, From, To, _, Distance),  
    Distance > 0.  
    
  @output("clean_routes").  
  @bind("clean_routes","databricks inCluster=true","","clean_air_routes").  
"""  

// Step 2: Graph analysis  
val analysisProgram = s"""  
  @bind("routes","databricks inCluster=true","","clean_air_routes").  
    
  edge(From, To, Dist) <- routes(From, To, Dist).  
  shortest_path(From, To, MinDist) <- #ASP(edge).  
    
  @output("shortest_path").  
  @bind("shortest_path","databricks inCluster=true","","flight_shortest_paths").  
"""  

// Execute both steps  
uk.co.prometheux.prometheuxreasoner.PrometheuxReasonerMain.main(Array(preprocessProgram))  
uk.co.prometheux.prometheuxreasoner.PrometheuxReasonerMain.main(Array(analysisProgram))  
```

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

### Automated Installation Script (Windows)

Source: https://docs.prometheux.ai/api/mcp/local

This PowerShell script automates the installation of the prometheux-mcp package on Windows systems. It handles dependency installation, prompts for credentials, and configures Claude Desktop.

```powershell
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/prometheuxresearch/px-mcp-server/main/install.ps1" -OutFile "install.ps1"
."/install.ps1"
```

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

### Querying relationships with SQL (Preview)

Source: https://docs.prometheux.ai/learn/vadalog/data-sources

Example of querying friendship relationships using SQL and previewing the first 5 results.

```VadaLog
@bind("friend_of_db", "neo4j username='neo4j', password='myPassw', host='neo4j-host', port=7680", "neo4j", "(:Person)-[:FRIEND_OF]->(:Person)").  

  

% Preview the first 5 friendship relationships  

friendship_preview() <- SELECT * FROM friend_of_db LIMIT 5.  

  

@output("friendship_preview").  


```
Tokens:
Raw