prescient 0.1.0 → 0.3.0

data/Rakefile

@@ -1,8 +1,11 @@
 # frozen_string_literal: true
 
 require "bundler/gem_tasks"
+require "rbconfig"
 require "rake/testtask"
 require "rubocop/rake_task"
+require "yard"
+require "yard/rake/yardoc_task"
 
 Rake::TestTask.new(:test) do |t|
   t.libs << "test"
@@ -10,21 +13,94 @@ Rake::TestTask.new(:test) do |t|
   t.test_files = FileList["test/**/*_test.rb"]
 end
 
-RuboCop::RakeTask.new
+RuboCop::RakeTask.new(:rubocop) do |task|
+  task.options = ["--parallel"]
+end
 
-desc "Run tests and linting"
-task default: %w[test rubocop]
+YARD::Rake::YardocTask.new(:yard)
+namespace :yard do
+  desc "Validate YARD documentation coverage"
+  task :validate do
+    require "open3"
+
+    stdout, stderr, status = Open3.capture3("bundle", "exec", "yard", "stats")
+    text = "#{stdout}\n#{stderr}"
+    puts text
+    abort("yard stats failed") unless status.success?
+
+    match = text.match(/([0-9]+(?:\.[0-9]+)?)%\s+documented/)
+    abort("Unable to determine YARD coverage") unless match
+
+    coverage = match[1].to_f
+    minimum = 99.0
+    if coverage < minimum
+      message = format(
+        "YARD coverage %<coverage>.2f%% is below %<minimum>.2f%%",
+        coverage: coverage,
+        minimum: minimum
+      )
+      abort(message)
+    end
 
-desc "Run tests with coverage"
-task :coverage do
-  ENV['COVERAGE'] = 'true'
-  Rake::Task[:test].execute
+    puts format("YARD coverage %.2f%%", coverage)
+  end
 end
 
+namespace :rbs do
+  desc "Remove generated RBS prototype files"
+  task :clobber do
+    sh "rm -rf tmp/sig"
+  end
+
+  desc "Generate disposable RBS prototypes into tmp/sig"
+  task :prototype do
+    sh "rm -rf tmp/sig"
+    sh "mkdir -p tmp/sig"
+    sh "bundle exec rbs prototype rb --out-dir=tmp/sig --base-dir=lib lib"
+  end
+
+  desc "Validate curated RBS signatures with Steep"
+  task :validate do
+    sh "bundle exec steep check"
+  end
+
+  desc "Open diff between curated and generated signatures"
+  task :diff do
+    sh "diff -ru sig tmp/sig || true"
+  end
+
+  desc "Generate disposable RBS prototypes and validate curated signatures"
+  task check: %i[prototype validate]
+end
+
+namespace :examples do
+  desc "Validate Ruby example syntax without contacting providers"
+  task :syntax do
+    Dir["examples/**/*.rb"].sort.each do |file|
+      sh RbConfig.ruby, "-c", file
+    end
+  end
+end
+
+desc "Validate GitHub Actions workflows"
+task :actionlint do
+  actionlint = if File.executable?(".tools/bin/actionlint")
+                 ".tools/bin/actionlint"
+               else
+                 ENV.fetch("ACTIONLINT", "actionlint")
+               end
+  sh actionlint
+end
+
+
+desc "Run tests and linting"
+task default: %w[test rubocop yard yard:validate rbs:validate examples:syntax]
+
 desc "Console with gem loaded"
 task :console do
   require "bundler/setup"
   require "prescient"
   require "irb"
+  ARGV.clear
   IRB.start
-end
+end

data/Steepfile

@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+target :prescient do
+  signature 'sig'
+  library 'json'
+  library 'net-http'
+
+  check 'lib/prescient/version.rb'
+  check 'lib/prescient/errors.rb'
+  check 'lib/prescient.rb'
+  check 'lib/prescient/client.rb'
+  check 'lib/prescient/base.rb'
+  check 'lib/prescient/provider/openai.rb'
+  check 'lib/prescient/provider/ollama.rb'
+  check 'lib/prescient/provider/anthropic.rb'
+  check 'lib/prescient/provider/huggingface.rb'
+end

data/VECTOR_SEARCH_GUIDE.md

@@ -2,23 +2,46 @@
 
 This guide provides a comprehensive overview of using Prescient with PostgreSQL's pgvector extension for semantic search and similarity matching.
 
+The runnable companion is [`examples/vector_search.rb`](examples/vector_search.rb);
+see the [examples guide](examples/README.md) for setup and the [main README](README.md)
+for the provider API.
+
 ## Quick Start
 
+For a reusable embedding store, use the library boundary rather than copying
+the example's application-specific SQL:
+
+```ruby
+require 'prescient'
+require 'pg'
+
+store = Prescient::Pgvector::Store.new(
+  connection: PG.connect(dbname: 'my_app'),
+  dimensions: 1536,
+)
+store.install!
+store.create_index!(metric: :cosine)
+```
+
+`Store` owns only its `prescient_embeddings` table. It does not manage document
+or chunk tables, connections, migrations outside that table, or the `pg` gem.
+Use `#upsert` and `#search` with embeddings of exactly the configured dimension.
+
 ### 1. Start Services
 
 ```bash
 # Start PostgreSQL with pgvector and Ollama
-docker-compose up -d postgres ollama
+docker compose up -d postgres ollama
 
 # Wait for services to be ready
-docker-compose logs -f postgres ollama
+docker compose logs -f postgres ollama
 ```
 
 ### 2. Initialize Models
 
 ```bash
 # Pull required Ollama models
-docker-compose up ollama-init
+docker compose run --rm ollama-init
 
 # Or manually:
 ./scripts/setup-ollama-models.sh
@@ -117,7 +140,7 @@ vector_str = "[#{embedding.join(',')}]"
 
 db.exec_params(
   "INSERT INTO document_embeddings (document_id, embedding_provider, embedding_model, embedding_dimensions, embedding, embedding_text) VALUES ($1, $2, $3, $4, $5, $6)",
-  [document_id, 'ollama', 'nomic-embed-text', 768, vector_str, content]
+  [document_id, 'ollama', 'nomic-embed-text', embedding.length, vector_str, content]
 )
 ```
 
@@ -130,17 +153,17 @@ query_embedding = client.generate_embedding(query_text)
 query_vector = "[#{query_embedding.join(',')}]"
 
 results = db.exec_params(
-  "SELECT d.title, d.content, de.embedding <=> $1::vector AS distance 
-   FROM documents d 
-   JOIN document_embeddings de ON d.id = de.document_id 
-   ORDER BY de.embedding <=> $1::vector 
+  "SELECT d.title, d.content, de.embedding <=> $1::vector AS distance
+   FROM documents d
+   JOIN document_embeddings de ON d.id = de.document_id
+   ORDER BY de.embedding <=> $1::vector
    LIMIT 5",
   [query_vector]
 )
 
 results.each do |row|
   similarity = 1 - row['distance'].to_f
-  puts "#{row['title']} (#{(similarity * 100).round(1)}% similar)"
+  puts "#{ row['title']} (#{ (similarity * 100).round(1)}% similar)"
 end
 ```
 
@@ -150,11 +173,11 @@ end
 # Search with metadata filtering
 results = db.exec_params(
   "SELECT d.title, de.embedding <=> $1::vector as distance
-   FROM documents d 
+   FROM documents d
    JOIN document_embeddings de ON d.id = de.document_id
    WHERE d.metadata->'tags' ? 'programming'
      AND d.metadata->>'difficulty' = 'beginner'
-   ORDER BY de.embedding <=> $1::vector 
+   ORDER BY de.embedding <=> $1::vector
    LIMIT 10",
   [query_vector]
 )
@@ -168,17 +191,17 @@ For large documents, split into chunks for better search granularity:
 def chunk_document(text, chunk_size: 500, overlap: 50)
   chunks = []
   start = 0
-  
+
   while start < text.length
     end_pos = [start + chunk_size, text.length].min
-    
+
     # Find word boundary to avoid cutting words
     if end_pos < text.length
       while end_pos > start && text[end_pos] != ' '
         end_pos -= 1
       end
     end
-    
+
     chunk = text[start...end_pos].strip
     chunks << {
       text: chunk,
@@ -186,11 +209,11 @@ def chunk_document(text, chunk_size: 500, overlap: 50)
       end_pos: end_pos,
       index: chunks.length
     }
-    
+
     start = end_pos - overlap
     break if start >= text.length
   end
-  
+
   chunks
 end
 
@@ -200,18 +223,18 @@ chunks.each do |chunk|
   # Insert chunk
   chunk_result = db.exec_params(
     "INSERT INTO document_chunks (document_id, chunk_index, chunk_text, chunk_metadata) VALUES ($1, $2, $3, $4) RETURNING id",
-    [document_id, chunk[:index], chunk[:text], {start_pos: chunk[:start_pos], end_pos: chunk[:end_pos]}.to_json]
+    [document_id, chunk[:index], chunk[:text], { start_pos: chunk[:start_pos], end_pos: chunk[:end_pos]}.to_json]
   )
   chunk_id = chunk_result[0]['id']
-  
+
   # Generate embedding for chunk
   chunk_embedding = client.generate_embedding(chunk[:text])
   chunk_vector = "[#{chunk_embedding.join(',')}]"
-  
+
   # Store chunk embedding
   db.exec_params(
     "INSERT INTO chunk_embeddings (chunk_id, document_id, embedding_provider, embedding_model, embedding_dimensions, embedding) VALUES ($1, $2, $3, $4, $5, $6)",
-    [chunk_id, document_id, 'ollama', 'nomic-embed-text', 768, chunk_vector]
+    [chunk_id, document_id, 'ollama', 'nomic-embed-text', chunk_embedding.length, chunk_vector]
   )
 end
 ```
@@ -224,20 +247,20 @@ For different dataset sizes and performance requirements:
 
 ```sql
 -- Small datasets (< 100K vectors): Fast build, good accuracy
-CREATE INDEX idx_embeddings_small 
-ON document_embeddings 
+CREATE INDEX idx_embeddings_small
+ON document_embeddings
 USING hnsw (embedding vector_cosine_ops)
 WITH (m = 8, ef_construction = 32);
 
 -- Medium datasets (100K - 1M vectors): Balanced
-CREATE INDEX idx_embeddings_medium 
-ON document_embeddings 
+CREATE INDEX idx_embeddings_medium
+ON document_embeddings
 USING hnsw (embedding vector_cosine_ops)
 WITH (m = 16, ef_construction = 64);
 
 -- Large datasets (> 1M vectors): High accuracy
-CREATE INDEX idx_embeddings_large 
-ON document_embeddings 
+CREATE INDEX idx_embeddings_large
+ON document_embeddings
 USING hnsw (embedding vector_cosine_ops)
 WITH (m = 32, ef_construction = 128);
 ```
@@ -251,9 +274,9 @@ SET hnsw.ef_search = 100;  -- Balanced (default)
 SET hnsw.ef_search = 200;  -- High accuracy, slower
 
 -- Monitor query performance
-EXPLAIN (ANALYZE, BUFFERS) 
-SELECT * FROM document_embeddings 
-ORDER BY embedding <=> '[0.1,0.2,...]'::vector 
+EXPLAIN (ANALYZE, BUFFERS)
+SELECT * FROM document_embeddings
+ORDER BY embedding <=> '[0.1,0.2,...]'::vector
 LIMIT 10;
 ```
 
@@ -268,7 +291,7 @@ texts.each_slice(10) do |batch|
   batch.each do |text|
     embedding = client.generate_embedding(text)
     embeddings << embedding
-    
+
     # Small delay to avoid rate limiting
     sleep(0.1)
   end
@@ -280,7 +303,7 @@ db.transaction do
     vector_str = "[#{embedding.join(',')}]"
     db.exec_params(
       "INSERT INTO document_embeddings (...) VALUES (...)",
-      [documents[index].id, 'ollama', 'nomic-embed-text', 768, vector_str, texts[index]]
+      [documents[index].id, 'ollama', 'nomic-embed-text', embedding.length, vector_str, texts[index]]
     )
   end
 end
@@ -295,13 +318,13 @@ Combine vector similarity with traditional text search:
 ```sql
 WITH vector_results AS (
   SELECT document_id, embedding <=> $1::vector as distance
-  FROM document_embeddings 
-  ORDER BY embedding <=> $1::vector 
+  FROM document_embeddings
+  ORDER BY embedding <=> $1::vector
   LIMIT 20
 ),
 text_results AS (
   SELECT id as document_id, ts_rank(to_tsvector(content), plainto_tsquery($2)) as rank
-  FROM documents 
+  FROM documents
   WHERE to_tsvector(content) @@ plainto_tsquery($2)
 )
 SELECT d.title, d.content,
@@ -322,16 +345,17 @@ Store embeddings from multiple providers for comparison:
 
 ```ruby
 providers = [
-  { client: Prescient.client(:ollama), name: 'ollama', model: 'nomic-embed-text', dims: 768 },
-  { client: Prescient.client(:openai), name: 'openai', model: 'text-embedding-3-small', dims: 1536 }
+  { client: Prescient.client(:ollama), name: 'ollama', model: 'nomic-embed-text' },
+  { client: Prescient.client(:openai), name: 'openai', model: 'text-embedding-3-small' }
 ]
 
 providers.each do |provider|
   next unless provider[:client].available?
-  
+
   embedding = provider[:client].generate_embedding(text)
+  provider[:dims] = embedding.length
   vector_str = "[#{embedding.join(',')}]"
-  
+
   db.exec_params(
     "INSERT INTO document_embeddings (document_id, embedding_provider, embedding_model, embedding_dimensions, embedding, embedding_text) VALUES ($1, $2, $3, $4, $5, $6)",
     [document_id, provider[:name], provider[:model], provider[:dims], vector_str, text]
@@ -348,14 +372,14 @@ end
 def track_search(query_text, results, provider, model)
   query_embedding = client.generate_embedding(query_text)
   query_vector = "[#{query_embedding.join(',')}]"
-  
+
   # Insert search query
   query_result = db.exec_params(
     "INSERT INTO search_queries (query_text, embedding_provider, embedding_model, query_embedding, result_count) VALUES ($1, $2, $3, $4, $5) RETURNING id",
     [query_text, provider, model, query_vector, results.length]
   )
   query_id = query_result[0]['id']
-  
+
   # Insert query results
   results.each_with_index do |result, index|
     db.exec_params(
@@ -371,14 +395,14 @@ end
 ```sql
 -- Popular search terms
 SELECT query_text, COUNT(*) as search_count
-FROM search_queries 
+FROM search_queries
 WHERE created_at > NOW() - INTERVAL '7 days'
 GROUP BY query_text
 ORDER BY search_count DESC
 LIMIT 10;
 
 -- Average similarity scores
-SELECT embedding_provider, embedding_model, 
+SELECT embedding_provider, embedding_model,
        AVG(similarity_score) as avg_similarity,
        COUNT(*) as result_count
 FROM query_results qr
@@ -400,11 +424,12 @@ ORDER BY hour;
 ### Common Issues
 
 **Slow queries:**
+
 ```sql
 -- Check if indexes are being used
-EXPLAIN (ANALYZE, BUFFERS) 
-SELECT * FROM document_embeddings 
-ORDER BY embedding <=> '[...]'::vector 
+EXPLAIN (ANALYZE, BUFFERS)
+SELECT * FROM document_embeddings
+ORDER BY embedding <=> '[...]'::vector
 LIMIT 10;
 
 -- Rebuild indexes if needed
@@ -412,10 +437,11 @@ REINDEX INDEX idx_document_embeddings_cosine;
 ```
 
 **Memory issues:**
+
 ```sql
 -- Check index sizes
 SELECT schemaname, tablename, indexname, pg_size_pretty(pg_relation_size(indexrelid)) as size
-FROM pg_stat_user_indexes 
+FROM pg_stat_user_indexes
 WHERE tablename LIKE '%embedding%'
 ORDER BY pg_relation_size(indexrelid) DESC;
 
@@ -424,6 +450,7 @@ SET work_mem = '256MB';
 ```
 
 **Dimension mismatches:**
+
 ```ruby
 # Validate embedding dimensions before storing
 expected_dims = 768
@@ -447,4 +474,4 @@ end
 - [pgvector Documentation](https://github.com/pgvector/pgvector)
 - [HNSW Algorithm](https://arxiv.org/abs/1603.09320)
 - [Vector Database Concepts](https://www.pinecone.io/learn/vector-database/)
-- [Embedding Best Practices](https://platform.openai.com/docs/guides/embeddings/what-are-embeddings)
+- [Embedding Best Practices](https://platform.openai.com/docs/guides/embeddings/what-are-embeddings)

data/docker-compose.yml

@@ -1,8 +1,6 @@
 # Docker Compose configuration for running Ollama with Prescient gem
 # This provides a local AI environment for development and testing
 
-version: '3.8'
-
 services:
   ollama:
     image: ollama/ollama:latest
@@ -65,7 +63,7 @@ services:
         # Pull chat model  
         curl -X POST http://ollama:11434/api/pull \
           -H "Content-Type: application/json" \
-          -d "{\"name\": \"llama3.1:8b\"}"
+          -d "{\"name\": \"llama3.2:3b\"}"
         
         echo "Models pulled successfully!"
       '
@@ -122,7 +120,7 @@ services:
       # Ollama configuration
       - OLLAMA_URL=http://ollama:11434
       - OLLAMA_EMBEDDING_MODEL=nomic-embed-text
-      - OLLAMA_CHAT_MODEL=llama3.1:8b
+      - OLLAMA_CHAT_MODEL=llama3.2:3b
       
       # Optional: Other AI provider configurations
       - OPENAI_API_KEY=${OPENAI_API_KEY:-}
@@ -150,4 +148,4 @@ volumes:
 
 networks:
   default:
-    name: prescient-network
+    name: prescient-network

data/examples/README.md

@@ -0,0 +1,45 @@
+# Prescient examples
+
+These scripts demonstrate the supported public API. They use the local
+library checkout, so run them from the repository root after installing the
+development dependencies:
+
+```bash
+bundle install
+```
+
+## Examples
+
+- `basic_usage.rb` — default-provider generation, embeddings, health checks,
+  provider selection, and custom configuration.
+- `custom_prompts.rb` — system prompts and no-context/with-context templates.
+- `custom_contexts.rb` — explicit context types, field matching, formatting,
+  and embedding field selection.
+- `vector_search.rb` — PostgreSQL/pgvector storage and similarity search.
+
+The first three examples use Ollama by default. Start Ollama and pull the
+current local models before running them:
+
+```bash
+docker compose up -d ollama
+docker compose run --rm ollama-init
+bundle exec ruby examples/basic_usage.rb
+```
+
+The vector-search example additionally requires PostgreSQL with pgvector:
+
+```bash
+docker compose up -d postgres ollama
+docker compose run --rm ollama-init
+DB_HOST=localhost bundle exec ruby examples/vector_search.rb
+```
+
+Cloud-provider examples require the corresponding credentials and provider
+configuration. The scripts are demonstrations rather than isolated test
+fixtures; they may make real provider requests when the configured service is
+available.
+
+See the [main README](../README.md) for configuration, fallback behavior,
+prompt templates, context exclusions, embeddings, and the public API. Rails
+applications can also use the [integration guide](../INTEGRATION_GUIDE.md),
+and PostgreSQL users should read the [pgvector guide](../VECTOR_SEARCH_GUIDE.md).

data/examples/basic_usage.rb

@@ -110,7 +110,7 @@ Prescient.configure do |config|
   config.add_provider(:custom_ollama, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     timeout: 60
   )
 end
@@ -120,4 +120,4 @@ puts "   Timeout: #{Prescient.configuration.timeout}s"
 puts "   Retry attempts: #{Prescient.configuration.retry_attempts}"
 puts "   Providers: #{Prescient.configuration.providers.keys.join(', ')}"
 
-puts "\n🎉 Examples completed!"
+puts "\n🎉 Examples completed!"

data/examples/custom_contexts.rb

@@ -16,7 +16,7 @@ Prescient.configure do |config|
   config.add_provider(:ecommerce, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     # Define your own context types - no hardcoded assumptions!
     context_configs: {
       'product' => {
@@ -88,7 +88,7 @@ Prescient.configure do |config|
   config.add_provider(:healthcare, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     context_configs: {
       'patient' => {
         fields: %w[name age gender medical_conditions medications],
@@ -158,7 +158,7 @@ Prescient.configure do |config|
   config.add_provider(:project_mgmt, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     context_configs: {
       'issue' => {
         fields: %w[title description status priority assignee labels created_date],
@@ -237,7 +237,7 @@ begin
     config.add_provider(:embedding_demo, Prescient::Provider::Ollama,
       url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
       embedding_model: 'nomic-embed-text',
-      chat_model: 'llama3.1:8b',
+      chat_model: 'llama3.2:3b',
       context_configs: {
         'blog_post' => {
           fields: %w[title content author tags category publish_date],
@@ -288,7 +288,7 @@ Prescient.configure do |config|
   config.add_provider(:no_config, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b'
+    chat_model: 'llama3.2:3b'
     # No context_configs defined - uses pure default behavior
   )
 end
@@ -352,4 +352,4 @@ puts "\n🎯 Best Practices:"
 puts "   - Define context_configs for your specific domain"
 puts "   - Use explicit 'type' field when context detection isn't reliable"
 puts "   - Exclude sensitive/metadata fields from embedding_fields"
-puts "   - Test with and without context configs to see the difference"
+puts "   - Test with and without context configs to see the difference"

data/examples/custom_prompts.rb

@@ -15,7 +15,7 @@ Prescient.configure do |config|
   config.add_provider(:customer_service, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     prompt_templates: {
       system_prompt: 'You are a friendly customer service representative. Be helpful, empathetic, and professional.',
       no_context_template: <<~TEMPLATE.strip,
@@ -73,7 +73,7 @@ Prescient.configure do |config|
   config.add_provider(:tech_docs, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     prompt_templates: {
       system_prompt: 'You are a technical documentation assistant. Provide clear, accurate, and detailed technical explanations with code examples when relevant.',
       no_context_template: <<~TEMPLATE.strip,
@@ -125,7 +125,7 @@ Prescient.configure do |config|
   config.add_provider(:creative, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     prompt_templates: {
       system_prompt: 'You are a creative writing assistant. Help with storytelling, character development, and creative inspiration. Be imaginative and encouraging.',
       no_context_template: <<~TEMPLATE.strip,
@@ -182,7 +182,7 @@ Prescient.configure do |config|
   config.add_provider(:custom_default, Prescient::Provider::Ollama,
     url: ENV.fetch('OLLAMA_URL', 'http://localhost:11434'),
     embedding_model: 'nomic-embed-text',
-    chat_model: 'llama3.1:8b',
+    chat_model: 'llama3.2:3b',
     prompt_templates: {
       # Only override the system prompt, keep default templates
       system_prompt: 'You are Sherlock Holmes. Approach every question with deductive reasoning and attention to detail.'
@@ -209,4 +209,4 @@ puts "\n💡 Tips:"
 puts "   - Use %{system_prompt}, %{query}, and %{context} placeholders in templates"
 puts "   - Templates use Ruby's % string formatting"
 puts "   - Override any or all template parts (system_prompt, no_context_template, with_context_template)"
-puts "   - Each provider can have completely different prompt behavior"
+puts "   - Each provider can have completely different prompt behavior"

data/examples/vector_search.rb

@@ -31,7 +31,7 @@ class VectorSearchExample
     
     # Check if services are available
     unless check_services_available
-      puts "❌ Required services not available. Please start with: docker-compose up -d"
+      puts "❌ Required services not available. Please start with: docker compose up -d"
       return
     end
 
@@ -327,4 +327,4 @@ puts "   - Try different embedding models (OpenAI, HuggingFace)"
 puts "   - Implement hybrid search (vector + keyword)"
 puts "   - Add document chunking for large texts"
 puts "   - Experiment with different similarity thresholds"
-puts "   - Add result re-ranking and filtering"
+puts "   - Add result re-ranking and filtering"