[examples] Improve the S3 Source example (#706)

dev/s3_upload.sh

@@ -18,11 +18,11 @@
 
 # Usage: ./minio-upload my-bucket my-file.zip
 
-bucket=$1
-file=$2
+host=$1
+url=$2
+bucket=$3
+file=$4
 
-host=localhost
-url=http://localhost:9000
 s3_key=minioadmin
 s3_secret=minioadmin
 

examples/applications/docker-chatbot/crawler.yaml

@@ -42,21 +42,12 @@ pipeline:
       allowed-domains: ["https://docs.langstream.ai"]
       forbidden-paths: []
       min-time-between-requests: 500
-      reindex-interval-seconds: 3600
       max-error-count: 5
       max-urls: 1000
       max-depth: 50
       handle-robots-file: true
-      user-agent: "" # this is computed automatically, but you can override it
       scan-html-documents: true
-      http-timeout: 10000
-      handle-cookies: true
-      max-unflushed-pages: 100
-      bucketName: "${secrets.s3.bucket-name}"
-      endpoint: "${secrets.s3.endpoint}"
-      access-key: "${secrets.s3.access-key}"
-      secret-key: "${secrets.s3.secret}"
-      region: "${secrets.s3.region}"
+      state-storage: disk
   - name: "Extract text"
     type: "text-extractor"
   - name: "Normalise text"

examples/applications/s3-source/README.md

@@ -0,0 +1,40 @@
+# Preprocessing Text 
+
+This sample application shows how to use some common NLP techniques to preprocess text data and then write the results to a Vector Database.
+
+We have two pipelines:
+
+The extract-text.yaml file defines a pipeline that will:
+
+- Extract text from document files (PDF, Word...)
+- Detect the language and filter out non-English documents
+- Normalize the text
+- Split the text into chunks
+- Write the chunks to a Vector Database, in this case DataStax Astra DB
+
+## Prerequisites
+
+Prepare some PDF files and upload them to a bucket in S3.
+
+## Deploy the LangStream application
+
+```
+./bin/langstream docker run test -app examples/applications/s3-source -s examples/secrets/secrets.yaml --docker-args="-p9900:9000"
+```
+
+Please note that here we are adding --docker-args="-p9900:9000" to expose the S3 API on port 9900.
+
+
+## Write some documents in the S3 bucket
+
+```
+# Upload a document to the S3 bucket
+dev/s3_upload.sh localhost http://localhost:9900 documents README.md
+dev/s3_upload.sh localhost http://localhost:9900 documents examples/applications/s3-source/simple.pdf
+```
+
+## Interact with the Chatbot
+
+Now you can use the developer UI to ask questions to the chatbot about your documents.
+
+If you have uploaded the README file then you should be able to ask "what is LangStream ?"

examples/applications/s3-source/chatbot.yaml

@@ -0,0 +1,101 @@
+#
+# Copyright DataStax, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+topics:
+  - name: "questions-topic"
+    creation-mode: create-if-not-exists
+  - name: "answers-topic"
+    creation-mode: create-if-not-exists
+  - name: "log-topic"
+    creation-mode: create-if-not-exists
+errors:
+    on-failure: "skip"
+pipeline:
+  - name: "convert-to-structure"
+    type: "document-to-json"
+    input: "questions-topic"
+    configuration:
+      text-field: "question"
+  - name: "compute-embeddings"
+    type: "compute-ai-embeddings"
+    configuration:
+      model: "${secrets.open-ai.embeddings-model}" # This needs to match the name of the model deployment, not the base model
+      embeddings-field: "value.question_embeddings"
+      text: "{{ value.question }}"
+      flush-interval: 0
+  - name: "lookup-related-documents"
+    type: "query-vector-db"
+    configuration:
+      datasource: "JdbcDatasource"
+      query: "SELECT text,embeddings_vector FROM documents ORDER BY cosine_similarity(embeddings_vector, CAST(? as FLOAT ARRAY)) DESC LIMIT 20"
+      fields:
+        - "value.question_embeddings"
+      output-field: "value.related_documents"
+  - name: "re-rank documents with MMR"
+    type: "re-rank"
+    configuration:
+      max: 5 # keep only the top 5 documents, because we have an hard limit on the prompt size
+      field: "value.related_documents"
+      query-text: "value.question"
+      query-embeddings: "value.question_embeddings"
+      output-field: "value.related_documents"
+      text-field: "record.text"
+      embeddings-field: "record.embeddings_vector"
+      algorithm: "MMR"
+      lambda: 0.5
+      k1: 1.2
+      b: 0.75
+  - name: "ai-chat-completions"
+    type: "ai-chat-completions"
+
+    configuration:
+      model: "${secrets.open-ai.chat-completions-model}" # This needs to be set to the model deployment name, not the base name
+      # on the log-topic we add a field with the answer
+      completion-field: "value.answer"
+      # we are also logging the prompt we sent to the LLM
+      log-field: "value.prompt"
+      # here we configure the streaming behavior
+      # as soon as the LLM answers with a chunk we send it to the answers-topic
+      stream-to-topic: "answers-topic"
+      # on the streaming answer we send the answer as whole message
+      # the 'value' syntax is used to refer to the whole value of the message
+      stream-response-completion-field: "value"
+      # we want to stream the answer as soon as we have 20 chunks
+      # in order to reduce latency for the first message the agent sends the first message
+      # with 1 chunk, then with 2 chunks....up to the min-chunks-per-message value
+      # eventually we want to send bigger messages to reduce the overhead of each message on the topic
+      min-chunks-per-message: 20
+      messages:
+        - role: system
+          content: |
+              An user is going to perform a questions, The documents below may help you in answering to their questions.
+              Please try to leverage them in your answer as much as possible.              
+              If you provide code or YAML snippets, please explicitly state that they are examples.
+              Do not provide information that is not related to the documents provided below.
+            
+              Documents:
+              {{# value.related_documents}}
+              {{ text}}
+              {{/ value.related_documents}}
+        - role: user
+          content: "{{ value.question}}"
+  - name: "cleanup-response"
+    type: "drop-fields"
+    output: "log-topic"
+    configuration:
+      fields:
+        - "question_embeddings"
+        - "related_documents"

examples/applications/s3-source/configuration.yaml

@@ -0,0 +1,38 @@
+#
+#
+# Copyright DataStax, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+configuration:
+  resources:
+    - type: "datasource"
+      name: "JdbcDatasource"
+      configuration:
+        service: "jdbc"
+        driverClass: "herddb.jdbc.Driver"
+        url: "${secrets.herddb.url}"
+        user: "${secrets.herddb.user}"
+        password: "${secrets.herddb.password}"
+    - type: "open-ai-configuration"
+      name: "OpenAI Azure configuration"
+      configuration:
+        url: "${secrets.open-ai.url}"
+        access-key: "${secrets.open-ai.access-key}"
+        provider: "${secrets.open-ai.provider}"
+  dependencies:
+    - name: "HerdDB.org JDBC Driver"
+      url: "https://repo1.maven.org/maven2/org/herddb/herddb-jdbc/0.28.0/herddb-jdbc-0.28.0-thin.jar"
+      sha512sum: "d8ea8fbb12eada8f860ed660cbc63d66659ab3506bc165c85c420889aa8a1dac53dab7906ef61c4415a038c5a034f0d75900543dd0013bdae50feafd46f51c8e"
+      type: "java-library"

examples/applications/text-processing/extract-text.yaml → examples/applications/s3-source/extract-text.yaml

@@ -15,9 +15,23 @@
 #
 
 name: "Extract and manipulate text"
-topics:
-  - name: "chunks-topic"
+assets:
+  - name: "documents-table"
+    asset-type: "jdbc-table"
     creation-mode: create-if-not-exists
+    config:
+      table-name: "documents"
+      datasource: "JdbcDatasource"
+      create-statements:
+        - |
+          CREATE TABLE documents (
+          filename TEXT,
+          chunk_id int,
+          num_tokens int,
+          lang TEXT,
+          text TEXT,
+          embeddings_vector FLOATA,
+          PRIMARY KEY (filename, chunk_id));
 pipeline:
   - name: "Read from S3"
     type: "s3-source"
@@ -27,7 +41,6 @@ pipeline:
       access-key: "${secrets.s3.access-key}"
       secret-key: "${secrets.s3.secret}"
       region: "${secrets.s3.region}"
-      idle-time: 5
   - name: "Extract text"
     type: "text-extractor"
   - name: "Normalise text"
@@ -73,10 +86,40 @@ pipeline:
   - name: "compute-embeddings"
     id: "step1"
     type: "compute-ai-embeddings"
-    output: "chunks-topic"
     configuration:
       model: "${secrets.open-ai.embeddings-model}" # This needs to match the name of the model deployment, not the base model
       embeddings-field: "value.embeddings_vector"
       text: "{{ value.text }}"
       batch-size: 10
       flush-interval: 500
+  - name: "Delete stale chunks"
+    type: "query"
+    configuration:
+      datasource: "JdbcDatasource"
+      when: "fn:toInt(properties.text_num_chunks) == (fn:toInt(properties.chunk_id) + 1)"
+      mode: "execute"
+      query: "DELETE FROM documents WHERE filename = ? AND chunk_id > ?"
+      output-field: "value.delete-results"
+      fields:
+        - "value.filename"
+        - "fn:toInt(value.chunk_id)"
+  - name: "Write"
+    type: "vector-db-sink"
+    configuration:
+      datasource: "JdbcDatasource"
+      table-name: "documents"
+      fields:
+        - name: "filename"
+          expression: "value.filename"
+          primary-key: true
+        - name: "chunk_id"
+          expression: "value.chunk_id"
+          primary-key: true
+        - name: "embeddings_vector"
+          expression: "fn:toListOfFloat(value.embeddings_vector)"
+        - name: "lang"
+          expression: "value.language"
+        - name: "text"
+          expression: "value.text"
+        - name: "num_tokens"
+          expression: "value.chunk_num_tokens"

examples/applications/text-processing/gateways.yaml → examples/applications/s3-source/gateways.yaml

@@ -16,6 +16,23 @@
 #
 
 gateways:
-  - id: "consume-chunks"
+  - id: "user-input"
+    type: produce
+    topic: "questions-topic"
+    parameters:
+      - sessionId
+    produceOptions:
+      headers:
+        - key: langstream-client-session-id
+          valueFromParameters: sessionId
+
+  - id: "bot-output"
     type: consume
-    topic: "chunks-topic"
+    topic: "answers-topic"
+    parameters:
+      - sessionId
+    consumeOptions:
+      filters:
+        headers:
+          - key: langstream-client-session-id
+            valueFromParameters: sessionId