import {z} from 'zod'

export type GetApiApiProjectIdTrainingKerasLearnIdArgs = {projectId: number, learnId: number};

export type GetApiApiProjectIdTrainingKerasLearnIdOptions = {baseUrl?: string | undefined, ApiKeyAuthentication?: string | undefined, fetch?: typeof fetch | undefined};

export const dependencyData = z.object({classes: z.array(z.string()), classesPerKey: z.record(z.string(), z.array(z.string())).optional(), blockNames: z.array(z.string()), featureCount: z.number().int(), sampleCount: z.number().int(), outputClasses: z.array(z.string())});

export const learnBlockType = z.enum(["anomaly", "anomaly-gmm", "keras", "keras-transfer-image", "keras-transfer-kws", "keras-object-detection", "keras-regression", "keras-akida", "keras-akida-transfer-image", "keras-akida-object-detection", "keras-visual-anomaly", "keras-freeform", "keras-self-contained-freeform", "keras-anomaly", "vlm"]);

export const kerasModelTypeEnum = z.enum(["int8", "float32", "akida", "requiresRetrain"]);

export const kerasVisualLayerType = z.enum(["dense", "conv1d", "conv2d", "reshape", "flatten", "dropout", "batchNormalization", "transfer_mobilenetv2_a35", "transfer_mobilenetv2_a1", "transfer_mobilenetv2_a05", "transfer_mobilenetv2_160_a1", "transfer_mobilenetv2_160_a75", "transfer_mobilenetv2_160_a5", "transfer_mobilenetv2_160_a35", "transfer_mobilenetv1_a25_d100", "transfer_mobilenetv1_a2_d100", "transfer_mobilenetv1_a1_d100", "transfer_kws_mobilenetv1_a1_d100", "transfer_kws_mobilenetv2_a35_d100", "transfer_kws_syntiant_ndp10x", "transfer_kws_conv2d_tiny", "object_ssd_mobilenet_v2_fpnlite_320x320", "fomo_mobilenet_v2_a01", "fomo_mobilenet_v2_a35", "transfer_organization", "transfer_akidanet_imagenet_160_a100", "transfer_akidanet_imagenet_160_a50", "transfer_akidanet_imagenet_160_a25", "transfer_akidanet_imagenet_224_a100", "transfer_akidanet_imagenet_224_a50", "transfer_akidanet_imagenet_224_a25", "fomo_akidanet_a50", "fomo_ad_gmm", "fomo_ad_patchcore"]);

export const kerasVisualLayer = z.object({type: kerasVisualLayerType, neurons: z.number().int().optional(), kernelSize: z.number().int().optional(), dropoutRate: z.number().optional(), columns: z.number().int().optional(), stack: z.number().int().optional(), enabled: z.boolean().optional(), organizationModelId: z.number().int().optional()});

export const augmentationPolicyImageEnum = z.enum(["none", "all"]);

export const augmentationPolicySpectrogram = z.object({enabled: z.boolean(), warping: z.boolean().optional(), freqMasking: z.enum(["none", "low", "high"]).optional(), timeMasking: z.enum(["none", "low", "high"]).optional(), gaussianNoise: z.enum(["none", "low", "high"]).optional()});

export const blockType = z.enum(["official", "personal", "enterprise", "community"]);

export const dSPGroupItem = z.object({name: z.string(), value: z.string().optional(), defaultValue: z.string(), type: z.string(), help: z.string().optional(), param: z.string(), selectOptions: z.array(z.object({value: z.string().optional(), selected: z.boolean().optional(), optionLabel: z.string().optional(), priority: z.number().optional(), romEstimate: z.number().optional(), needsOps: z.array(z.string()).optional(), needsFeatures: z.array(z.string()).optional()})).optional(), readonly: z.boolean(), shouldShow: z.boolean(), showIf: z.object({parameter: z.string(), operator: z.enum(["eq", "neq"]), value: z.string()}).optional(), invalidText: z.string().optional(), section: z.enum(["advanced", "augmentation", "modelProfiling"]).optional(), multiline: z.boolean().optional(), required: z.boolean(), hint: z.string().optional(), placeholder: z.string().optional(), showClickToSet: z.boolean(), valid: z.array(z.object({})).optional(), items: z.object({}).optional(), properties: z.object({}).optional(), minVal: z.number().optional(), maxVal: z.number().optional()});

export const blockDisplayCategory = z.enum(["classical", "tao", "developer-preview"]);

export const transferLearningModel = z.object({name: z.string(), shortName: z.string(), abbreviatedName: z.string().optional(), description: z.string(), hasNeurons: z.boolean(), hasDropout: z.boolean(), defaultNeurons: z.number().int().optional(), defaultDropout: z.number().optional(), defaultLearningRate: z.number().optional(), defaultTrainingCycles: z.number().optional(), hasImageAugmentation: z.boolean().optional(), type: kerasVisualLayerType, learnBlockType: learnBlockType, organizationModelId: z.number().int().optional(), implementationVersion: z.number().int().optional(), repositoryUrl: z.string().optional(), customBlockRef: z.string().optional(), author: z.string(), blockType: blockType, customParameters: z.array(dSPGroupItem).optional(), displayCategory: blockDisplayCategory.optional(), blockNoLongerAvailable: z.object({reasonMarkdown: z.string(), reasonHtml: z.string()}).optional(), hasExpertMode: z.boolean()});

export const akidaEdgeLearningConfig = z.object({enabled: z.boolean(), additionalClasses: z.number().optional(), neuronsPerClass: z.number().optional()});

export const anomalyCapacity = z.enum(["low", "medium", "high"]);

export const kerasModelVariantEnum = z.enum(["int8", "float32", "akida"]);

export const blockParamsVisualAnomalyPatchcore = z.object({backbone: z.string().optional(), numLayers: z.number().int().optional(), poolSize: z.number().int().optional(), samplingRatio: z.number().optional(), numNearestNeighbors: z.number().int().optional()});

export const blockParamsVisualAnomalyGmm = z.object({backbone: z.string().optional()});

export const blockParameters = z.union([blockParamsVisualAnomalyPatchcore, blockParamsVisualAnomalyGmm]);

export const blockThreshold = z.object({key: z.string(), description: z.string(), helpText: z.string(), suggestedValue: z.number().optional(), suggestedValueText: z.string().optional(), value: z.union([z.number(), z.string()]), dropdownOptions: z.array(z.object({description: z.string(), value: z.string()})).optional()});

export const kFoldsValidationConfig = z.object({numFolds: z.number(), includeFinalTrainingRun: z.boolean().optional(), extraEpochs: z.number().optional()});

export const kerasResponse = z.object({success: z.boolean(), error: z.string().optional(), dependencies: dependencyData, trained: z.boolean(), name: z.string(), type: learnBlockType.optional(), script: z.string(), minimumConfidenceRating: z.number(), selectedModelType: kerasModelTypeEnum, mode: z.enum(["visual", "expert"]), visualLayers: z.array(kerasVisualLayer), trainingCycles: z.number().int(), learningRate: z.number(), batchSize: z.number().int().optional(), defaultBatchSize: z.number().int(), shape: z.string(), trainTestSplit: z.number().optional(), autoClassWeights: z.boolean().optional(), useLearnedOptimizer: z.boolean().optional(), augmentationPolicyImage: augmentationPolicyImageEnum, augmentationPolicySpectrogram: augmentationPolicySpectrogram.optional(), transferLearningModels: z.array(transferLearningModel), profileInt8: z.boolean(), skipEmbeddingsAndMemory: z.boolean(), akidaEdgeLearningConfig: akidaEdgeLearningConfig.optional(), customValidationMetadataKey: z.string().optional(), showAdvancedTrainingSettings: z.boolean(), showAugmentationTrainingSettings: z.boolean(), customParameters: z.record(z.string(), z.string().nullable()).optional(), anomalyCapacity: anomalyCapacity.optional(), lastShownModelVariant: kerasModelVariantEnum.optional(), blockParameters: blockParameters.optional(), thresholds: z.array(blockThreshold), kFoldsValidationConfig: kFoldsValidationConfig.optional(), isMultiHead: z.boolean().optional()});

export const validateResponse = async <Output>(
  schema: {
    readonly '~standard': {
      readonly validate: (value: unknown) => ({ readonly value: Output; readonly issues?: undefined } | { readonly issues: ReadonlyArray<{ readonly message: string; readonly path?: ReadonlyArray<PropertyKey | { readonly key: PropertyKey }> | undefined }> }) | Promise<{ readonly value: Output; readonly issues?: undefined } | { readonly issues: ReadonlyArray<{ readonly message: string; readonly path?: ReadonlyArray<PropertyKey | { readonly key: PropertyKey }> | undefined }> }>
    }
  },
  value: unknown
): Promise<Output> => {
  const result = await schema['~standard'].validate(value)

  if (result.issues !== undefined) {
    const detail = result.issues
      .map(({ message, path }) =>
        path?.length
          ? `${path.map((segment) => String(typeof segment === 'object' ? segment.key : segment)).join('.')}: ${message}`
          : message
      )
      .join('; ')

    throw new Error(`Response does not match the schema: ${detail}`, { cause: { issues: result.issues } })
  }

  return result.value
};

/**
 * Keras information
 * 
 * GET /api/{projectId}/training/keras/{learnId}
 * 
 * Also accepts auth: JWTAuthentication, JWTHttpHeaderAuthentication, OAuth2
 */
export const getApiApiProjectIdTrainingKerasLearnId = async (args: GetApiApiProjectIdTrainingKerasLearnIdArgs, options?: GetApiApiProjectIdTrainingKerasLearnIdOptions) => {
const {baseUrl = "https://studio.edgeimpulse.com/v1", fetch: fetchFn = fetch, ApiKeyAuthentication} = options ?? {}

const query = new URLSearchParams()
const headers = new Headers()
headers.set('accept', "application/json")

if (ApiKeyAuthentication !== undefined) headers.set("x-api-key", ApiKeyAuthentication)

const url = `${baseUrl.replace(/\/+$/, '')}/api/${encodeURIComponent(String(args.projectId))}/training/keras/${encodeURIComponent(String(args.learnId))}${query.size > 0 ? `?${query}` : ''}`

const res = await fetchFn(url, {method: "GET", headers})

if (!res.ok) throw new Error(`Request failed with status ${res.status}`, {cause: {status: res.status, body: await res.text()}})

return validateResponse(kerasResponse, await res.json())
};