Config file improvements (drasi-project#55)

* Remove platform examples

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* Fix trading example configuration schema to match API naming conventions

The API uses a mix of snake_case and camelCase depending on the struct:
- SourceConfig/ReactionConfig top-level fields: snake_case (auto_start, bootstrap_provider)
- Nested config DTOs (QueryConfigDto, SseReactionConfigDto, etc.): camelCase (sourceId, ssePath)

Changes:
- YAML: auto_start, bootstrap_provider for SourceConfig fields
- DrasiClient.ts: sourceId for query sources, ssePath/heartbeatIntervalMs for SSE config
- DrasiClient.ts: autoStart for QueryConfigDto, auto_start for ReactionConfig

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Remove deleted platform example tests from validation suite

Removes test references to examples that were deleted in commit 7ffee84:
- debug-platform-building-comfort
- drasi-platform

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Update trading example readme file.

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* Add CONTRIBUTING.md with learning exercises and contribution ideas

Includes:
- 9 learning exercises (beginner to advanced) for understanding Drasi
- Prioritized feature contribution suggestions
- Documentation improvement ideas
- Getting started guide for contributors

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* remove unused plugin config models

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* Add wide range of config file examples

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* Add two-pass config validation to catch unknown and snake_case fields

Implements field validation that runs before typed deserialization to
catch typos and snake_case field names that would otherwise be silently
ignored. This prevents configuration errors from going unnoticed.

Key changes:
- Add src/config/validation.rs with field validation for all config types
- Integrate validation into load_config_file() as first-pass check
- Add rename_all_fields="camelCase" to SourceConfig/ReactionConfig enums
- Fix test files and example configs to use camelCase consistently
- Add comprehensive test coverage for parsing failures

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Fix mock source dataType and query label mismatches in examples

The mock source only supports "sensor", "counter", and "generic" data
types. The examples incorrectly used "sensor_live" which falls through
to generic. Also fixed query labels to match actual generated data:
- dataType "sensor" generates SensorReading nodes (not Sensor)
- dataType "counter" generates Counter nodes
- dataType "generic" generates Generic nodes

Added unit tests to verify YAML deserialization of mock source config.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Fix templates in example config files

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* Add Handlebars template syntax validation to config validation

Uses Handlebars' built-in register_template_string() method to validate
template syntax at config parse time, catching errors early rather than
at runtime when reactions try to execute templates.

Validates templates in:
- Log reaction: defaultTemplate and routes templates
- SSE reaction: defaultTemplate and routes templates
- HTTP reaction: body and url fields in routes

Catches common template errors:
- Unclosed braces: {{after.Name
- Empty expressions: {{}}
- Unclosed blocks: {{#if condition}}...
- Mismatched helpers: {{#each}}...{{/if}}

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Replace manual field validation with custom serde deserializers

- Add custom Deserialize implementations for SourceConfig and ReactionConfig
  that extract common fields (kind, id, autoStart, queries) manually and
  delegate remaining fields to inner DTOs with deny_unknown_fields

- Add #[serde(deny_unknown_fields)] to all source and reaction inner DTOs
  to reject unknown/snake_case fields at deserialization time

- Inline adaptive batch fields in HttpAdaptiveReactionConfigDto and
  GrpcAdaptiveReactionConfigDto to avoid flatten + deny_unknown_fields conflict

- Remove ~1100 lines of hardcoded field validation from validation.rs
  (SOURCE_COMMON_FIELDS, source_fields(), reaction_fields(), etc.)

- Add 44 unit tests for custom deserializers covering valid configs,
  missing fields, unknown fields, snake_case rejection, and error messages

- Fix trading example client to use camelCase field names (autoStart,
  ssePath, heartbeatIntervalMs) instead of snake_case

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

* Fixing Getting Started integration test

Signed-off-by: Agent of Reality <agentx@agentofreality.com>

* refactor: unify config validation using deny_unknown_fields

- Add deny_unknown_fields to DrasiServerConfig, DrasiLibInstanceConfig,
  QueryConfigDto, and SourceSubscriptionConfigDto
- Implement custom deserializer for StateStoreConfig matching the pattern
  used by SourceConfig and ReactionConfig
- Create RedbStateStoreConfigDto with deny_unknown_fields for strict
  field validation
- Simplify validation.rs to only handle Handlebars template syntax
  validation (remove manual field arrays)
- Remove SERVER_FIELDS, INSTANCE_FIELDS, QUERY_FIELDS arrays
- Add comprehensive tests for StateStoreConfig validation

This change makes all config types use consistent validation:
- Simple structs use #[serde(deny_unknown_fields)]
- Tagged enums (Source, Reaction, StateStore) use custom deserializers
  with inner DTOs that have deny_unknown_fields

Unknown fields are now rejected during deserialization instead of
requiring separate manual validation.

* refactor: move BootstrapProviderConfig to drasi-server and standardize on 'kind' discriminator

Changes:
- Created src/api/models/bootstrap.rs with local BootstrapProviderConfig
  - Custom deserializer with deny_unknown_fields on all inner DTOs
  - Conversion implementations to drasi-lib types
  - Comprehensive test coverage (20 tests)
- Changed discriminator field from 'type' to 'kind' for consistency
  with SourceConfig, ReactionConfig, and StateStoreConfig
- Removed unused AdaptiveBatchConfigDto struct (dead code)
- Updated factories.rs to use local BootstrapProviderConfig
- Updated init/prompts.rs to use local types
- Updated all example configs and tests

This completes the config validation consistency work:
- All tagged enums now use 'kind' as discriminator
- All inner DTOs have deny_unknown_fields
- All custom deserializers provide contextual error messages

* fix: rename bootstrapProvider type to kind in config

* docs: update config examples to use camelCase and correct field names

- Change bootstrap provider discriminator from 'type' to 'kind'
- Convert all snake_case field names to camelCase in YAML examples
- Update README.md, CLAUDE.md, and examples/getting-started/README.md
- Ensures documentation matches current config schema validation

* test: cleanup obsolete tests and add CLI test coverage

- Add validate_command_test.rs (19 tests) for CLI validate command
- Add init_output_test.rs (18 tests) for init output validation
- Rename test files to use _test.rs suffix convention
- Delete obsolete manual tests (grpc/, http/, sse-console/)
- Delete broken shell scripts referencing missing files
- Update tests/README.md with comprehensive documentation
- Fix config/server-docker.yaml to use current schema format

Total test count: 486 automated tests

---------

Signed-off-by: Agent of Reality <agentx@agentofreality.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: agentofreality

CLAUDE.md

@@ -89,70 +89,70 @@ id: "my-server"  # Unique server ID (defaults to UUID if not specified)
 # Server settings
 host: "0.0.0.0"
 port: 8080
-log_level: "info"
-persist_config: true  # Enable persistence (default)
-persist_index: false  # Use RocksDB for persistent indexing (default: false, uses in-memory)
+logLevel: "info"
+persistConfig: true  # Enable persistence (default)
+persistIndex: false  # Use RocksDB for persistent indexing (default: false, uses in-memory)
 
 # Optional state store for plugin state persistence
-# state_store:
+# stateStore:
 #   kind: redb
 #   path: ./data/state.redb
 
 # Optional capacity defaults (cascades to queries/reactions)
 # Supports environment variables like other fields
-# default_priority_queue_capacity: 10000
-# default_priority_queue_capacity: "${PRIORITY_QUEUE_CAPACITY:-10000}"
-# default_dispatch_buffer_capacity: 1000
-# default_dispatch_buffer_capacity: "${DISPATCH_BUFFER_CAPACITY:-1000}"
+# defaultPriorityQueueCapacity: 10000
+# defaultPriorityQueueCapacity: "${PRIORITY_QUEUE_CAPACITY:-10000}"
+# defaultDispatchBufferCapacity: 1000
+# defaultDispatchBufferCapacity: "${DISPATCH_BUFFER_CAPACITY:-1000}"
 
 # Sources (parsed into plugin instances)
 sources:
   - kind: mock
     id: "sensors"
-    auto_start: true
+    autoStart: true
 
 # Queries
 queries:
   - id: "high-temp"
     query: "MATCH (s:Sensor) WHERE s.temperature > 75 RETURN s"
     queryLanguage: Cypher
     sources:
-      - source_id: "sensors"
-    auto_start: true
+      - sourceId: "sensors"
+    autoStart: true
 
 # Reactions
 reactions:
   - kind: log
     id: "log-temps"
     queries:
       - "high-temp"
-    auto_start: true
+    autoStart: true
 ```
 
 For multiple DrasiLib instances, use the `instances` array (legacy single-instance fields continue to work and map to the first instance):
 
 ```yaml
 host: "0.0.0.0"
 port: 8080
-log_level: "info"
-persist_config: true
+logLevel: "info"
+persistConfig: true
 
 instances:
   - id: "analytics"
-    persist_index: true
-    state_store:
+    persistIndex: true
+    stateStore:
       kind: redb
       path: ./data/analytics-state.redb
     sources:
       - kind: mock
         id: "sensors"
-        auto_start: true
+        autoStart: true
     queries:
       - id: "high-temp"
         query: "MATCH (s:Sensor) WHERE s.temperature > 75 RETURN s"
         queryLanguage: Cypher
         sources:
-          - source_id: "sensors"
+          - sourceId: "sensors"
   - id: "monitoring"
     sources: []
     queries: []
@@ -173,12 +173,12 @@ DrasiServer separates two independent concepts:
 **Persistence is enabled when:**
 - Config file is provided on startup (`--config path/to/config.yaml`)
 - Config file is writable
-- `persist_config: true` in server settings (default)
+- `persistConfig: true` in server settings (default)
 
 **Persistence is disabled when:**
 - No config file provided (server starts with empty configuration)
 - Config file is read-only
-- `persist_config: false` in server settings
+- `persistConfig: false` in server settings
 
 **Read-Only mode is enabled ONLY when:**
 - Config file is not writable (file permissions prevent writing)

Cargo.lock

@@ -87,6 +87,7 @@ indexmap = "2.0"
 thiserror = "1.0"
 dotenvy = "0.15"
 inquire = "0.7"
+handlebars = "6"
 
 [dev-dependencies]
 # Testing utilities