Observability for Chroma with Highlight

Step 1

Install the Traceloop SDK and initialize it. It will automatically log all calls to Chroma using its SDK - both query calls and index calls.

import chromadb
from traceloop.sdk import Traceloop
from traceloop.sdk.decorators import workflow

Traceloop.init()

chroma_client = chromadb.Client()
collection = chroma_client.create_collection(
    name="my_collection", embedding_function=embedding_function
)

claim_query_result = collection.query(
  query_texts=query, include=["documents", "distances"], n_results=3
 )

import * as traceloop from "@traceloop/node-server-sdk";
import { ChromaClient, IncludeEnum, OpenAIEmbeddingFunction } from "chromadb";
import OpenAI from "openai";
import fs from "fs";
import { ChatCompletionMessageParam } from "openai/resources/chat/completions";

const openai_client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });
const client = new ChromaClient();

traceloop.initialize({
  appName: "sample_chromadb",
  apiKey: process.env.TRACELOOP_API_KEY,
  disableBatch: true,
});

const embeddingFunction = new OpenAIEmbeddingFunction({
  openai_api_key: process.env.OPENAI_API_KEY ?? "",
});

const scifactCorpusCollection = client.getOrCreateCollection({
  name: "scifact_corpus",
  embeddingFunction,
});

const claimData = fs
  .readFileSync("data/scifact/scifact_claims.jsonl")
  .toString()
  .split("\n")
  .map((each) => {
    try {
      return JSON.parse(each);
    } catch (e) {
      // Continue
    }
  });
const corpusData = fs
  .readFileSync("data/scifact/scifact_corpus.jsonl")
  .toString()
  .split("\n")
  .map((each) => {
    try {
      return JSON.parse(each);
    } catch (e) {
      // Continue
    }
  });

const batchSize = 100;

const buildPromptWithContext = (
  claim: any,
  context: any,
): ChatCompletionMessageParam[] => [
  {
    role: "system",
    content:
      "I will ask you to assess whether a particular scientific claim, based on evidence provided. " +
      "Output only the text 'True' if the claim is true, 'False' if the claim is false, or 'NEE' if there's " +
      "not enough evidence.",
  },
  {
    role: "user",
    content: `
The evidence is the following:

${context.join(" ")}

Assess the following claim on the basis of the evidence. Output only the text 'True' if the claim is true,
'False' if the claim is false, or 'NEE' if there's not enough evidence. Do not output any other text.

Claim:
${claim}

Assessment:
`,
  },
];

async function assessClaims(claims: any) {
  const claimQueryResult = await (
    await scifactCorpusCollection
  ).query({
    queryTexts: claims,
    include: [IncludeEnum.Documents, IncludeEnum.Distances],
    nResults: 3,
  });
  const responses = [];

  for (let i = 0; i < claimQueryResult.documents.length; i++) {
    const claim = claims[i];
    const context = claimQueryResult.documents[i];
    if (context.length === 0) {
      responses.push("NEE");
      continue;
    }

    const response = await openai_client.chat.completions.create({
      model: "gpt-3.5-turbo",
      messages: buildPromptWithContext(claim, context),
      max_tokens: 3,
    });

    const formattedResponse = response.choices[0].message.content?.replace(
      "., ",
      "",
    );
    console.log("Claim: ", claim);
    console.log("Response: ", formattedResponse);
    responses.push(formattedResponse);
  }

  return responses;
}

traceloop.withAssociationProperties({}, async () => {
  const samples = claimData.slice(0, 2); // Get a sample of 2 claims
  assessClaims(samples.map((sample) => sample["claim"]));
});

Step 2

Route traces to Highlights’s OTLP endpoint and set the Highlight project ID in the headers.

TRACELOOP_BASE_URL=https://otel.highlight.io:4318
TRACELOOP_HEADERS="x-highlight-project=<YOUR_HIGHLIGHT_PROJECT_ID>"

Docs

Observability for Chroma with Highlight

Step 1

Step 2

Discover use cases

Trace Indexing pipelines with Chroma

Trace your RAG retrieval pipeline

More things you can connect to {platform}

OpenAI

Cohere

Ollama

Gemini

Watsonx

Bedrock

Anthropic

LangChain

LlamaIndex

Mistral AI

together.ai