Add WhisperX configuration contract

This commit is contained in:
2026-05-03 15:23:47 -05:00
parent 598584cf2d
commit c2a715d3f6
8 changed files with 227 additions and 10 deletions

View File

@@ -11,6 +11,11 @@ This repository currently contains a **working scaffold** with strict config loa
- `pipeline.yml`: pipeline/workspace settings (`workspace`, `storage`, `whisperx`, `seriatim`, `audita`, `analyzer`, `notification`) - `pipeline.yml`: pipeline/workspace settings (`workspace`, `storage`, `whisperx`, `seriatim`, `audita`, `analyzer`, `notification`)
- `session.yml`: per-session settings (`session_id`, `inputs`, optional metadata) - `session.yml`: per-session settings (`session_id`, `inputs`, optional metadata)
WhisperX config contract in `pipeline.yml`:
- required: `whisperx.transcribe_url`
- defaulted when omitted: `whisperx.language` (`en`), `whisperx.timeout` (`30m`), `whisperx.retries` (`3`), `whisperx.retry_delay` (`2s`), `whisperx.concurrency` (`2`)
Decoding is strict (`KnownFields(true)`), so unknown YAML fields fail fast. Decoding is strict (`KnownFields(true)`), so unknown YAML fields fail fast.
Example minimal files are available under `examples/`: Example minimal files are available under `examples/`:

View File

@@ -102,10 +102,15 @@ Key behavior:
- Strict YAML decoding via `yaml.Decoder.KnownFields(true)`. - Strict YAML decoding via `yaml.Decoder.KnownFields(true)`.
- Unknown fields are rejected. - Unknown fields are rejected.
- Combined resolved config type keeps source paths (`PipelinePath`, `SessionPath`) for provenance/errors. - Combined resolved config type keeps source paths (`PipelinePath`, `SessionPath`) for provenance/errors.
- WhisperX optional fields are defaulted during load for deterministic resolved config values.
Validation currently enforces: Validation currently enforces:
- `pipeline.workspace.root` is required. - `pipeline.workspace.root` is required.
- `pipeline.whisperx.transcribe_url` is required and must be a valid URL.
- `pipeline.whisperx.timeout` and `pipeline.whisperx.retry_delay` must parse as Go durations.
- `pipeline.whisperx.retries` must be `>= 0`.
- `pipeline.whisperx.concurrency` must be `> 0`.
- `session.session_id` is required. - `session.session_id` is required.
- `session.inputs.speakers_file`, `autocorrect_file`, `glossary_file` are required. - `session.inputs.speakers_file`, `autocorrect_file`, `glossary_file` are required.
- At least one audio source: `session.inputs.audio_dir` or non-empty `session.inputs.audio_files`. - At least one audio source: `session.inputs.audio_dir` or non-empty `session.inputs.audio_files`.

View File

@@ -5,7 +5,12 @@ storage:
backend: local backend: local
whisperx: whisperx:
timeout: 15m transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
language: en
timeout: 30m
retries: 3
retry_delay: 2s
concurrency: 2
seriatim: seriatim:
timeout: 30s timeout: 30s
@@ -20,4 +25,3 @@ analyzer:
notification: notification:
timeout: 10s timeout: 10s

View File

@@ -145,7 +145,7 @@ func writeValidConfigFiles(t *testing.T, workspaceRoot string) (string, string)
storage: storage:
backend: s3 backend: s3
whisperx: whisperx:
timeout: 15m transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
seriatim: seriatim:
timeout: 30s timeout: 30s
audita: audita:

View File

@@ -42,9 +42,12 @@ type StorageConfig struct {
// WhisperXConfig configures WhisperX adapter settings. // WhisperXConfig configures WhisperX adapter settings.
type WhisperXConfig struct { type WhisperXConfig struct {
BaseURL string `yaml:"base_url"` TranscribeURL string `yaml:"transcribe_url"`
Language string `yaml:"language"`
Timeout string `yaml:"timeout"` Timeout string `yaml:"timeout"`
Concurrency int `yaml:"concurrency"` Retries *int `yaml:"retries"`
RetryDelay string `yaml:"retry_delay"`
Concurrency *int `yaml:"concurrency"`
} }
// SeriatimConfig configures seriatim adapter settings. // SeriatimConfig configures seriatim adapter settings.

View File

@@ -15,6 +15,7 @@ func LoadPipeline(path string) (*PipelineConfig, error) {
if err := decodeStrictYAML("pipeline", path, &cfg); err != nil { if err := decodeStrictYAML("pipeline", path, &cfg); err != nil {
return nil, fmt.Errorf("load pipeline config: %w", err) return nil, fmt.Errorf("load pipeline config: %w", err)
} }
applyPipelineDefaults(&cfg)
return &cfg, nil return &cfg, nil
} }
@@ -75,3 +76,36 @@ func shortName(path, fallback string) string {
} }
return base return base
} }
func applyPipelineDefaults(cfg *PipelineConfig) {
if cfg == nil {
return
}
applyWhisperXDefaults(&cfg.WhisperX)
}
func applyWhisperXDefaults(cfg *WhisperXConfig) {
if cfg == nil {
return
}
if cfg.Language == "" {
cfg.Language = "en"
}
if cfg.Timeout == "" {
cfg.Timeout = "30m"
}
if cfg.RetryDelay == "" {
cfg.RetryDelay = "2s"
}
if cfg.Concurrency == nil {
cfg.Concurrency = intPtr(2)
}
if cfg.Retries == nil {
cfg.Retries = intPtr(3)
}
}
func intPtr(v int) *int {
p := v
return &p
}

View File

@@ -14,13 +14,14 @@ func TestLoadAndValidate(t *testing.T) {
sessionYAML string sessionYAML string
wantLoadErr string wantLoadErr string
wantValidate string wantValidate string
checkDefault bool
}{ }{
{ {
name: "valid minimal config", name: "valid minimal config",
pipelineYAML: `workspace: pipelineYAML: `workspace:
root: /tmp/narratio root: /tmp/narratio
whisperx: whisperx:
timeout: 10m transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
seriatim: seriatim:
timeout: 30s timeout: 30s
audita: audita:
@@ -37,6 +38,7 @@ inputs:
autocorrect_file: ./autocorrect.yml autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml glossary_file: ./glossary.yml
`, `,
checkDefault: true,
}, },
{ {
name: "unknown pipeline field fails", name: "unknown pipeline field fails",
@@ -53,6 +55,23 @@ inputs:
`, `,
wantLoadErr: "pipeline file", wantLoadErr: "pipeline file",
}, },
{
name: "unknown whisperx field fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
bogus: true
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantLoadErr: "strict decode failed",
},
{ {
name: "unknown session field fails", name: "unknown session field fails",
pipelineYAML: `workspace: pipelineYAML: `workspace:
@@ -72,6 +91,8 @@ unknown_field: true
name: "missing required field fails", name: "missing required field fails",
pipelineYAML: `workspace: pipelineYAML: `workspace:
root: /tmp/narratio root: /tmp/narratio
whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
`, `,
sessionYAML: `session_id: "" sessionYAML: `session_id: ""
inputs: inputs:
@@ -82,11 +103,43 @@ inputs:
`, `,
wantValidate: "session config \"session.yml\" invalid: session.session_id is required", wantValidate: "session config \"session.yml\" invalid: session.session_id is required",
}, },
{
name: "missing transcribe_url fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.transcribe_url is required",
},
{
name: "invalid transcribe_url fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
transcribe_url: ://
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.transcribe_url must be a valid URL",
},
{ {
name: "invalid timeout fails", name: "invalid timeout fails",
pipelineYAML: `workspace: pipelineYAML: `workspace:
root: /tmp/narratio root: /tmp/narratio
whisperx: whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
timeout: definitely-not-a-duration timeout: definitely-not-a-duration
`, `,
sessionYAML: `session_id: 2026-05-03 sessionYAML: `session_id: 2026-05-03
@@ -98,6 +151,57 @@ inputs:
`, `,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.timeout must be a valid duration", wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.timeout must be a valid duration",
}, },
{
name: "invalid retry_delay fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
retry_delay: not-a-duration
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.retry_delay must be a valid duration",
},
{
name: "negative retries fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
retries: -1
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.retries must be >= 0",
},
{
name: "invalid concurrency fails",
pipelineYAML: `workspace:
root: /tmp/narratio
whisperx:
transcribe_url: https://transcription.ai.rakestrawhome.com/transcribe
concurrency: 0
`,
sessionYAML: `session_id: 2026-05-03
inputs:
audio_dir: ./audio
speakers_file: ./speakers.yml
autocorrect_file: ./autocorrect.yml
glossary_file: ./glossary.yml
`,
wantValidate: "pipeline config \"pipeline.yml\" invalid: pipeline.whisperx.concurrency must be > 0",
},
} }
for _, tt := range tests { for _, tt := range tests {
@@ -127,6 +231,23 @@ inputs:
if cfg.SessionPath != sessionPath { if cfg.SessionPath != sessionPath {
t.Fatalf("SessionPath = %q, want %q", cfg.SessionPath, sessionPath) t.Fatalf("SessionPath = %q, want %q", cfg.SessionPath, sessionPath)
} }
if tt.checkDefault {
if cfg.Pipeline.WhisperX.Language != "en" {
t.Fatalf("whisperx.language = %q, want %q", cfg.Pipeline.WhisperX.Language, "en")
}
if cfg.Pipeline.WhisperX.Timeout != "30m" {
t.Fatalf("whisperx.timeout = %q, want %q", cfg.Pipeline.WhisperX.Timeout, "30m")
}
if cfg.Pipeline.WhisperX.RetryDelay != "2s" {
t.Fatalf("whisperx.retry_delay = %q, want %q", cfg.Pipeline.WhisperX.RetryDelay, "2s")
}
if cfg.Pipeline.WhisperX.Retries == nil || *cfg.Pipeline.WhisperX.Retries != 3 {
t.Fatalf("whisperx.retries = %v, want 3", cfg.Pipeline.WhisperX.Retries)
}
if cfg.Pipeline.WhisperX.Concurrency == nil || *cfg.Pipeline.WhisperX.Concurrency != 2 {
t.Fatalf("whisperx.concurrency = %v, want 2", cfg.Pipeline.WhisperX.Concurrency)
}
}
err = Validate(cfg) err = Validate(cfg)
if tt.wantValidate != "" { if tt.wantValidate != "" {
@@ -146,8 +267,20 @@ inputs:
} }
func TestValidateMissingAudioSource(t *testing.T) { func TestValidateMissingAudioSource(t *testing.T) {
retries := 3
concurrency := 2
cfg := &Config{ cfg := &Config{
Pipeline: &PipelineConfig{Workspace: WorkspaceConfig{Root: "/tmp/narratio"}}, Pipeline: &PipelineConfig{
Workspace: WorkspaceConfig{Root: "/tmp/narratio"},
WhisperX: WhisperXConfig{
TranscribeURL: "https://transcription.ai.rakestrawhome.com/transcribe",
Language: "en",
Timeout: "30m",
Retries: &retries,
RetryDelay: "2s",
Concurrency: &concurrency,
},
},
Session: &SessionConfig{ Session: &SessionConfig{
SessionID: "2026-05-03", SessionID: "2026-05-03",
Inputs: SessionInputsConfig{ Inputs: SessionInputsConfig{

View File

@@ -2,6 +2,7 @@ package config
import ( import (
"fmt" "fmt"
"net/url"
"strings" "strings"
"time" "time"
) )
@@ -32,10 +33,10 @@ func validatePipeline(cfg *PipelineConfig) error {
if strings.TrimSpace(cfg.Workspace.Root) == "" { if strings.TrimSpace(cfg.Workspace.Root) == "" {
return fmt.Errorf("pipeline.workspace.root is required") return fmt.Errorf("pipeline.workspace.root is required")
} }
if err := validateWhisperX(cfg.WhisperX); err != nil {
if err := validateDuration("pipeline.whisperx.timeout", cfg.WhisperX.Timeout); err != nil {
return err return err
} }
if err := validateDuration("pipeline.seriatim.timeout", cfg.Seriatim.Timeout); err != nil { if err := validateDuration("pipeline.seriatim.timeout", cfg.Seriatim.Timeout); err != nil {
return err return err
} }
@@ -52,6 +53,38 @@ func validatePipeline(cfg *PipelineConfig) error {
return nil return nil
} }
func validateWhisperX(cfg WhisperXConfig) error {
if strings.TrimSpace(cfg.TranscribeURL) == "" {
return fmt.Errorf("pipeline.whisperx.transcribe_url is required")
}
u, err := url.Parse(cfg.TranscribeURL)
if err != nil || u.Scheme == "" || u.Host == "" {
if err != nil {
return fmt.Errorf("pipeline.whisperx.transcribe_url must be a valid URL: %w", err)
}
return fmt.Errorf("pipeline.whisperx.transcribe_url must be a valid URL")
}
if err := validateDuration("pipeline.whisperx.timeout", cfg.Timeout); err != nil {
return err
}
if err := validateDuration("pipeline.whisperx.retry_delay", cfg.RetryDelay); err != nil {
return err
}
if cfg.Retries == nil {
return fmt.Errorf("pipeline.whisperx.retries must be set (defaults should populate this)")
}
if *cfg.Retries < 0 {
return fmt.Errorf("pipeline.whisperx.retries must be >= 0")
}
if cfg.Concurrency == nil {
return fmt.Errorf("pipeline.whisperx.concurrency must be set (defaults should populate this)")
}
if *cfg.Concurrency <= 0 {
return fmt.Errorf("pipeline.whisperx.concurrency must be > 0")
}
return nil
}
func validateSession(cfg *SessionConfig) error { func validateSession(cfg *SessionConfig) error {
if strings.TrimSpace(cfg.SessionID) == "" { if strings.TrimSpace(cfg.SessionID) == "" {
return fmt.Errorf("session.session_id is required") return fmt.Errorf("session.session_id is required")