Improve collection and query loading; fix reflection panic for certain faker methods that require arguments

Author: daniel-percona

README.md

@@ -157,7 +157,38 @@ The additional collection and query definitions can be found here:
 * [collections](./resources/collections/)
 * [queries](./resources/queries/)
 
-### 4. Docker & Kubernetes
+
+### 4. Workload Configuration & Loading
+
+You can supply your own collections and queries using the `PLGM_COLLECTIONS_PATH` and `PLGM_QUERIES_PATH` environment variables (or the corresponding config file fields). 
+
+plgm supports two loading modes:
+
+#### 1. Single File Mode
+If you point to a specific file, plgm will load **only** that file, regardless of its name.
+
+```bash
+# Loads only my_custom_workload.json
+export PLGM_COLLECTIONS_PATH="./resources/collections/my_custom_workload.json"
+```
+
+#### 2. Directory Mode (Multi-file)
+If you point to a folder, plgm will scan and merge **all** `.json` files found in that folder. This allows you to split complex schemas across multiple files.
+
+```bash
+# Loads all .json files in the /custom folder
+export PLGM_COLLECTIONS_PATH="./resources/custom_collections/"
+```
+
+#### Default Workload Filtering
+When using **Directory Mode**, the behavior depends on the `PLGM_DEFAULT_WORKLOAD` setting:
+
+* **`true` (Default):** Loads **all** JSON files in the directory, including `default.json`.
+* **`false` (Custom):** Loads all JSON files **except** `default.json`. 
+  * *Use Case:* You can keep the example `default.json` in your folder for reference but exclude it from your actual test run by setting `PLGM_DEFAULT_WORKLOAD=false`.
+
+
+### 5. Docker & Kubernetes
 Prefer running in a container? We have a dedicated guide for building Docker images and running performance jobs directly inside Kubernetes (recommended for accurate network latency testing).
 
 [View the Docker & Kubernetes Guide](docker.md)

internal/config/collections.go

@@ -43,7 +43,13 @@ type CollectionsFile struct {
 	Collections []CollectionDefinition `json:"collections"`
 }
 
-// LoadCollections filters files based on the 'loadDefault' flag (from DefaultWorkload config).
+// LoadCollections loads files from a path.
+// If path is a folder:
+//   - If loadDefault is TRUE: Loads ALL .json files (including default.json).
+//   - If loadDefault is FALSE: Loads ALL .json files EXCEPT default.json.
+//
+// If path is a file:
+//   - Loads the file unconditionally.
 func LoadCollections(path string, loadDefault bool) (*CollectionsFile, error) {
 	if path == "" {
 		return &CollectionsFile{}, nil
@@ -67,11 +73,11 @@ func LoadCollections(path string, loadDefault bool) (*CollectionsFile, error) {
 			}
 
 			isDefault := strings.EqualFold(entry.Name(), "default.json")
-			if loadDefault && !isDefault {
-				continue // Skip non-default files
-			}
+
+			// If default_workload is FALSE, we explicitly ignore 'default.json' in folder mode.
+			// If default_workload is TRUE, we load everything (no continue).
 			if !loadDefault && isDefault {
-				continue // Skip default file
+				continue
 			}
 
 			fullPath := filepath.Join(path, entry.Name())
@@ -82,25 +88,12 @@ func LoadCollections(path string, loadDefault bool) (*CollectionsFile, error) {
 			allCollections = append(allCollections, loaded.Collections...)
 		}
 	} else {
-		// Single file: apply filtering logic to the specific file
-		filename := filepath.Base(path)
-		isDefault := strings.EqualFold(filename, "default.json")
-
-		shouldLoad := true
-		if loadDefault && !isDefault {
-			shouldLoad = false
-		}
-		if !loadDefault && isDefault {
-			shouldLoad = false
-		}
-
-		if shouldLoad {
-			loaded, err := loadCollectionsFromFile(path)
-			if err != nil {
-				return nil, err
-			}
-			allCollections = append(allCollections, loaded.Collections...)
+		// Single file: Always load it if the user specifically pointed to it.
+		loaded, err := loadCollectionsFromFile(path)
+		if err != nil {
+			return nil, err
 		}
+		allCollections = append(allCollections, loaded.Collections...)
 	}
 
 	return &CollectionsFile{Collections: allCollections}, nil

internal/config/queries.go

@@ -24,7 +24,13 @@ type QueriesFile struct {
 	Queries []QueryDefinition
 }
 
-// LoadQueries filters files based on the 'loadDefault' flag.
+// LoadQueries loads files from a path.
+// If path is a folder:
+//   - If loadDefault is TRUE: Loads ALL .json files (including default.json).
+//   - If loadDefault is FALSE: Loads ALL .json files EXCEPT default.json.
+//
+// If path is a file:
+//   - Loads the file unconditionally.
 func LoadQueries(path string, loadDefault bool) (*QueriesFile, error) {
 	if path == "" {
 		return &QueriesFile{}, nil
@@ -48,9 +54,9 @@ func LoadQueries(path string, loadDefault bool) (*QueriesFile, error) {
 			}
 
 			isDefault := strings.EqualFold(entry.Name(), "default.json")
-			if loadDefault && !isDefault {
-				continue
-			}
+
+			// If default_workload is FALSE, we explicitly ignore 'default.json' in folder mode.
+			// If default_workload is TRUE, we load everything.
 			if !loadDefault && isDefault {
 				continue
 			}
@@ -63,24 +69,12 @@ func LoadQueries(path string, loadDefault bool) (*QueriesFile, error) {
 			allQueries = append(allQueries, loaded.Queries...)
 		}
 	} else {
-		filename := filepath.Base(path)
-		isDefault := strings.EqualFold(filename, "default.json")
-
-		shouldLoad := true
-		if loadDefault && !isDefault {
-			shouldLoad = false
-		}
-		if !loadDefault && isDefault {
-			shouldLoad = false
-		}
-
-		if shouldLoad {
-			loaded, err := loadQueriesFromFile(path)
-			if err != nil {
-				return nil, err
-			}
-			allQueries = append(allQueries, loaded.Queries...)
+		// Single file: Always load it.
+		loaded, err := loadQueriesFromFile(path)
+		if err != nil {
+			return nil, err
 		}
+		allQueries = append(allQueries, loaded.Queries...)
 	}
 
 	return &QueriesFile{Queries: allQueries}, nil

internal/datagen/random.go

@@ -38,13 +38,36 @@ func RandomValueWithFaker(def config.CollectionField, faker *gofakeit.Faker) int
 		method := fakerVal.MethodByName(methodName)
 
 		if method.IsValid() {
-			results := method.Call(nil)
-			if len(results) > 0 {
-				return results[0].Interface()
+			// Check argument count to prevent panics
+			numArgs := method.Type().NumIn()
+
+			if numArgs == 0 {
+				// e.g. Name(), Email() - No args required
+				results := method.Call(nil)
+				if len(results) > 0 {
+					return results[0].Interface()
+				}
+			} else if numArgs == 1 && method.Type().In(0).Kind() == reflect.Int {
+				// e.g. Sentence(wordCount) or Paragraph(paragraphCount)
+				// Smart logic: Use configuration constraints if available
+				argVal := 5 // Default fallback
+
+				if def.MaxLength > 0 {
+					argVal = def.MaxLength
+				} else if def.ArraySize > 0 {
+					argVal = def.ArraySize
+				}
+
+				results := method.Call([]reflect.Value{reflect.ValueOf(argVal)})
+				if len(results) > 0 {
+					return results[0].Interface()
+				}
 			}
+			// Note: Methods requiring >1 args or non-int args are skipped here
+			// and will fall through to the switch/default below.
 		}
 
-		// Fallback for special providers
+		// Fallback for special providers (handling cases reflection might miss or specific overrides)
 		switch strings.ToLower(def.Provider) {
 		case "uuid":
 			return faker.UUID()
@@ -91,7 +114,8 @@ func RandomValueWithFaker(def config.CollectionField, faker *gofakeit.Faker) int
 		if def.Provider == "" {
 			return fmt.Sprintf("str-%d", rng.Intn(100000))
 		}
-		return "val" // Should be handled by provider logic
+		// If provider was invalid or skipped in reflection, we land here.
+		return "val"
 
 	case "bool", "boolean":
 		return rng.Intn(2) == 0