Files
LocalAI/core/http/routes/ollama.go
Richard Palethorpe 3fa7b2955c feat(pii): NER tier engine — privacy-filter.cpp backend + NER-centric PII filter (#10360)
Squashed feat/pii-ner-tier-engine rebased onto master (was 45 commits; see
backup/pii-ner-tier-engine-prerebase). Net change:

- privacy-filter.cpp: standalone GGML engine for the openai-privacy-filter
  PII/NER token classifier, wired as a LocalAI gRPC backend (CPU/CUDA/Vulkan).
  TokenClassify moves off the patched llama.cpp path onto this backend.
- PII filter reworked to be NER-centric (encoder/NER detection tier scanning
  whole conversations as one document), with a recreated bounded restricted-
  regex secret-matching pattern detector tier alongside it (per-model
  pii_detection.builtins / .patterns + core/services/routing/piipattern).
- Detection labelled by source (ner vs pattern); backend trace / confidence /
  debug observability; analyze/redact exposed as a synchronous API.
- Instance-wide default detector policy + per-usecase default-on; request
  filtering extended to completions, embeddings, edits & Ollama.
- React UI: NER-centric PII editor, detector-models table, pattern/builtins
  editor, middleware default-policy UI.
- Gallery: privacy-filter-multilingual token-classify model + NER install
  filter; token_classify known_usecase; batch sized to context for NER models.
  privacy-filter backend registered in the backend gallery (cpu/vulkan/cuda-13
  meta + image entries with a capabilities map) matching its CI matrix jobs,
  and an /import-model auto-detect importer (PrivacyFilterImporter, narrow
  privacy-filter GGUF detection) replacing the prior pref-only registration.

Reconciled against master's independent evolution:

- Dropped master's PIIPatternOverrides feature (global-pattern runtime
  overrides + /api/pii/patterns API + runtime_settings.json persistence). The
  per-model NER + pattern-detector design supersedes it; it was built on the
  global redactor pattern set this branch replaced.
- Reverted the llama.cpp Score carry-patch (0006-server-task-type-score):
  removed the patch and restored master's grpc-server.cpp Score RPC (direct
  llama_decode, slot-loop bypass) and LLAMA_VERSION pin, plus master's
  model_config validation forbidding score + chat/completion/embeddings on
  llama-cpp. token_classify is unaffected (it runs on the privacy-filter
  backend, not llama-cpp).

Assisted-by: Claude:claude-opus-4-8 [Claude Code]

Signed-off-by: Richard Palethorpe <io@richiejp.com>
2026-06-18 11:45:22 +01:00

178 lines
6.9 KiB
Go

package routes
import (
"context"
"github.com/google/uuid"
"github.com/labstack/echo/v4"
"github.com/mudler/LocalAI/core/application"
"github.com/mudler/LocalAI/core/config"
"github.com/mudler/LocalAI/core/http/endpoints/ollama"
"github.com/mudler/LocalAI/core/http/middleware"
"github.com/mudler/LocalAI/core/schema"
"github.com/mudler/LocalAI/core/services/routing/pii"
"github.com/mudler/LocalAI/core/services/routing/piiadapter"
"github.com/mudler/LocalAI/pkg/distributedhdr"
)
func RegisterOllamaRoutes(app *echo.Echo,
re *middleware.RequestExtractor,
application *application.Application,
) {
traceMiddleware := middleware.TraceMiddleware(application)
usageMiddleware := middleware.UsageMiddleware(application.StatsRecorder(), application.FallbackUser())
nodeHeaderMiddleware := middleware.ExposeNodeHeader(application.ApplicationConfig())
// Chat endpoint: POST /api/chat
chatHandler := ollama.ChatEndpoint(
application.ModelConfigLoader(),
application.ModelLoader(),
application.TemplatesEvaluator(),
application.ApplicationConfig(),
)
chatMiddleware := []echo.MiddlewareFunc{
nodeHeaderMiddleware,
usageMiddleware,
traceMiddleware,
re.BuildFilteredFirstAvailableDefaultModel(config.BuildUsecaseFilterFn(config.FLAG_CHAT)),
re.SetModelAndConfig(func() schema.LocalAIRequest { return new(schema.OllamaChatRequest) }),
setOllamaChatRequestContext(application.ApplicationConfig()),
pii.RequestMiddleware(application.PIIRedactor(), application.PIIEvents(), piiadapter.OllamaChat(), application.FallbackUser(), pii.WithNERResolver(application.PIINERResolver()), pii.WithPolicyResolver(application.PIIPolicyResolver())),
}
app.POST("/api/chat", chatHandler, chatMiddleware...)
// Generate endpoint: POST /api/generate
generateHandler := ollama.GenerateEndpoint(
application.ModelConfigLoader(),
application.ModelLoader(),
application.TemplatesEvaluator(),
application.ApplicationConfig(),
)
generateMiddleware := []echo.MiddlewareFunc{
nodeHeaderMiddleware,
usageMiddleware,
traceMiddleware,
re.BuildFilteredFirstAvailableDefaultModel(config.BuildUsecaseFilterFn(config.FLAG_CHAT)),
re.SetModelAndConfig(func() schema.LocalAIRequest { return new(schema.OllamaGenerateRequest) }),
setOllamaGenerateRequestContext(application.ApplicationConfig()),
pii.RequestMiddleware(application.PIIRedactor(), application.PIIEvents(), piiadapter.OllamaGenerate(), application.FallbackUser(), pii.WithNERResolver(application.PIINERResolver()), pii.WithPolicyResolver(application.PIIPolicyResolver())),
}
app.POST("/api/generate", generateHandler, generateMiddleware...)
// Embed endpoints: POST /api/embed and /api/embeddings
embedHandler := ollama.EmbedEndpoint(
application.ModelConfigLoader(),
application.ModelLoader(),
application.ApplicationConfig(),
)
embedMiddleware := []echo.MiddlewareFunc{
nodeHeaderMiddleware,
usageMiddleware,
traceMiddleware,
re.BuildFilteredFirstAvailableDefaultModel(config.BuildUsecaseFilterFn(config.FLAG_EMBEDDINGS)),
re.SetModelAndConfig(func() schema.LocalAIRequest { return new(schema.OllamaEmbedRequest) }),
pii.RequestMiddleware(application.PIIRedactor(), application.PIIEvents(), piiadapter.OllamaEmbed(), application.FallbackUser(), pii.WithNERResolver(application.PIINERResolver()), pii.WithPolicyResolver(application.PIIPolicyResolver())),
}
app.POST("/api/embed", embedHandler, embedMiddleware...)
app.POST("/api/embeddings", embedHandler, embedMiddleware...)
// Model management endpoints (no model-specific middleware needed)
app.GET("/api/tags", ollama.ListModelsEndpoint(application.ModelConfigLoader(), application.ModelLoader()))
app.HEAD("/api/tags", ollama.ListModelsEndpoint(application.ModelConfigLoader(), application.ModelLoader()))
app.POST("/api/show", ollama.ShowModelEndpoint(application.ModelConfigLoader()))
app.GET("/api/ps", ollama.ListRunningEndpoint(application.ModelConfigLoader(), application.ModelLoader()))
app.GET("/api/version", ollama.VersionEndpoint())
app.HEAD("/api/version", ollama.VersionEndpoint())
}
// RegisterOllamaRootEndpoint registers the Ollama "/" health check.
// This is separate because it conflicts with the web UI and is gated behind a CLI flag.
func RegisterOllamaRootEndpoint(app *echo.Echo) {
app.GET("/", ollama.HeartbeatEndpoint())
app.HEAD("/", ollama.HeartbeatEndpoint())
}
// setOllamaChatRequestContext sets up context and cancellation for Ollama chat requests
func setOllamaChatRequestContext(appConfig *config.ApplicationConfig) echo.MiddlewareFunc {
return func(next echo.HandlerFunc) echo.HandlerFunc {
return func(c echo.Context) error {
input, ok := c.Get(middleware.CONTEXT_LOCALS_KEY_LOCALAI_REQUEST).(*schema.OllamaChatRequest)
if !ok || input.Model == "" {
return echo.ErrBadRequest
}
cfg, ok := c.Get(middleware.CONTEXT_LOCALS_KEY_MODEL_CONFIG).(*config.ModelConfig)
if !ok || cfg == nil {
return echo.ErrBadRequest
}
correlationID := uuid.New().String()
c.Response().Header().Set("X-Correlation-ID", correlationID)
reqCtx := c.Request().Context()
c1, cancel := context.WithCancel(appConfig.Context)
stop := context.AfterFunc(reqCtx, cancel)
defer func() {
stop()
cancel()
}()
ctxWithCorrelationID := context.WithValue(c1, middleware.CorrelationIDKey, correlationID)
ctxWithCorrelationID = distributedhdr.Inherit(ctxWithCorrelationID, reqCtx)
input.Context = ctxWithCorrelationID
input.Cancel = cancel
if cfg.Model == "" {
cfg.Model = input.Model
}
c.Set(middleware.CONTEXT_LOCALS_KEY_LOCALAI_REQUEST, input)
c.Set(middleware.CONTEXT_LOCALS_KEY_MODEL_CONFIG, cfg)
return next(c)
}
}
}
// setOllamaGenerateRequestContext sets up context and cancellation for Ollama generate requests
func setOllamaGenerateRequestContext(appConfig *config.ApplicationConfig) echo.MiddlewareFunc {
return func(next echo.HandlerFunc) echo.HandlerFunc {
return func(c echo.Context) error {
input, ok := c.Get(middleware.CONTEXT_LOCALS_KEY_LOCALAI_REQUEST).(*schema.OllamaGenerateRequest)
if !ok || input.Model == "" {
return echo.ErrBadRequest
}
cfg, ok := c.Get(middleware.CONTEXT_LOCALS_KEY_MODEL_CONFIG).(*config.ModelConfig)
if !ok || cfg == nil {
return echo.ErrBadRequest
}
correlationID := uuid.New().String()
c.Response().Header().Set("X-Correlation-ID", correlationID)
reqCtx := c.Request().Context()
c1, cancel := context.WithCancel(appConfig.Context)
stop := context.AfterFunc(reqCtx, cancel)
defer func() {
stop()
cancel()
}()
ctxWithCorrelationID := context.WithValue(c1, middleware.CorrelationIDKey, correlationID)
ctxWithCorrelationID = distributedhdr.Inherit(ctxWithCorrelationID, reqCtx)
input.Ctx = ctxWithCorrelationID
input.Cancel = cancel
if cfg.Model == "" {
cfg.Model = input.Model
}
c.Set(middleware.CONTEXT_LOCALS_KEY_LOCALAI_REQUEST, input)
c.Set(middleware.CONTEXT_LOCALS_KEY_MODEL_CONFIG, cfg)
return next(c)
}
}
}