cohere

package main

import (
	"context"
	"fmt"

	"github.com/conduitio/conduit-commons/config"
	"github.com/conduitio/conduit-commons/opencdc"
	sdk "github.com/conduitio/conduit-processor-sdk"
	"github.com/conduitio/conduit/pkg/foundation/log"
	"github.com/conduitio/conduit/pkg/plugin/processor/builtin/internal/exampleutil"
)

func main() {
	p := NewCommandProcessor(log.Nop())
	p.client = &mockCommandClient{}

	exampleutil.RunExample(p, exampleutil.Example{
		Summary: `Generate responses using Cohere's command model`,
		Description: `
This example demonstrates how to use the Cohere command processor to generate responses for a record's ` + "`.Payload.After`" + ` field.
The processor sends the input text to the Cohere API and replaces it with the model's response.`,
		Config: config.Config{
			commandProcessorConfigApiKey: "apikey",
			commandProcessorConfigPrompt: "hello",
		},
		Have: opencdc.Record{
			Operation: opencdc.OperationUpdate,
			Position:  opencdc.Position("pos-1"),
			Payload: opencdc.Change{
				After: opencdc.RawData("who are you?"),
			},
		},
		Want: sdk.SingleRecord{
			Operation: opencdc.OperationUpdate,
			Position:  opencdc.Position("pos-1"),
			Payload: opencdc.Change{
				After: opencdc.RawData("cohere command response content"),
			},
		},
	})

}

type mockCommandClient struct{}

func (m mockCommandClient) command(ctx context.Context, content string) (string, error) {
	if content == "" {
		return "", fmt.Errorf("mocked api error")
	}
	return "cohere command response content", nil
}

Output:

processor transformed record:
--- before
+++ after
@@ -1,10 +1,10 @@
 {
   "position": "cG9zLTE=",
   "operation": "update",
   "metadata": null,
   "key": null,
   "payload": {
     "before": null,
-    "after": "who are you?"
+    "after": "cohere command response content"
   }
 }

p := NewEmbedProcessor(log.Nop())
p.client = mockEmbedClient{}

exampleutil.RunExample(p, exampleutil.Example{
	Summary: `Generate embeddings using Cohere's embedding model`,
	Description: `
This example demonstrates how to use the Cohere embedding processor to generate embeddings for a record.
The processor extracts text from the specified input field (default: ".Payload.After"), sends it to the Cohere API,
and stores the resulting embeddings in the record's ".Payload.After" field as compressed data using the zstd algorithm.

In this example, the processor is configured with a mock client and an API key. The input record's metadata is updated
to include the embedding model used ("embed-english-v2.0"). Note that the compressed embeddings cannot be directly compared
in this test, so the focus is on verifying the metadata update.`,
	Config: config.Config{
		"apiKey": "fake-api-key",
	},
	Have: opencdc.Record{
		Operation: opencdc.OperationCreate,
		Position:  opencdc.Position("pos-1"),
		Metadata:  map[string]string{},
	},
	Want: sdk.SingleRecord{
		Operation: opencdc.OperationCreate,
		Position:  opencdc.Position("pos-1"),
		Metadata:  opencdc.Metadata{"cohere.embed.model": "embed-english-v2.0"},
	},
})

Output:

processor transformed record:
--- before
+++ after
@@ -1,10 +1,12 @@
 {
   "position": "cG9zLTE=",
   "operation": "create",
-  "metadata": {},
+  "metadata": {
+    "cohere.embed.model": "embed-english-v2.0"
+  },
   "key": null,
   "payload": {
     "before": null,
     "after": null
   }
 }

package main

import (
	"context"
	"fmt"
	"math"

	"github.com/conduitio/conduit-commons/config"
	"github.com/conduitio/conduit-commons/opencdc"
	sdk "github.com/conduitio/conduit-processor-sdk"
	"github.com/conduitio/conduit/pkg/foundation/log"
	"github.com/conduitio/conduit/pkg/plugin/processor/builtin/internal/exampleutil"
)

func main() {
	p := NewRerankProcessor(log.Nop())
	p.client = &mockRerankClient{}

	exampleutil.RunExample(p, exampleutil.Example{
		Summary: `Generate responses using Cohere's rerank model`,
		Description: `
This example demonstrates how to use the Cohere rerank processor.This takes in a query and a list of texts and produces an ordered 
array with each text assigned a relevance score. The processor extracts text from the specified input field (default: ".Payload.After"), 
sends it to the Cohere API, and stores the response in the record's ".Payload.After" field.

In this example, the processor is configured with a mock client and an API key. The input record's metadata is updated
to include the rerank model used ("rerank-v3.5").`,
		Config: config.Config{
			"apiKey": "fakeapiKey",
			"query":  "What is the capital of the United States?",
		},
		Have: opencdc.Record{
			Operation: opencdc.OperationUpdate,
			Position:  opencdc.Position("pos-1"),
			Payload: opencdc.Change{
				After: opencdc.RawData("Carson City is the capital city of the American state of Nevada."),
			},
		},
		Want: sdk.SingleRecord{
			Operation: opencdc.OperationUpdate,
			Position:  opencdc.Position("pos-1"),
			Payload: opencdc.Change{
				After: opencdc.RawData(`{"document":{"text":"Carson City is the capital city of the American state of Nevada."},"index":0,"relevance_score":0.9}`),
			},
		},
	})

}

type mockRerankClient struct{}

func (m mockRerankClient) rerank(ctx context.Context, docs []string) ([]RerankResult, error) {
	if len(docs) == 0 {
		return nil, fmt.Errorf("mocked api error")
	}
	result := make([]RerankResult, 0, len(docs))

	mockedScore := 1.0
	for i, d := range docs {
		mockedScore -= 0.1
		mockedScore = math.Round(mockedScore*10) / 10
		res := RerankResult{
			Index:          i,
			RelevanceScore: mockedScore,
		}
		res.Document.Text = d
		result = append(result, res)
	}

	return result, nil
}

Output:

processor transformed record:
--- before
+++ after
@@ -1,10 +1,10 @@
 {
   "position": "cG9zLTE=",
   "operation": "update",
   "metadata": null,
   "key": null,
   "payload": {
     "before": null,
-    "after": "Carson City is the capital city of the American state of Nevada."
+    "after": "{\"document\":{\"text\":\"Carson City is the capital city of the American state of Nevada.\"},\"index\":0,\"relevance_score\":0.9}"
   }
 }