mchades commented on code in PR #12900: URL: https://github.com/apache/gravitino/pull/12900#discussion_r4254481452
########## web-v2/web/src/lib/utils/semanticModel.js: ########## @@ -0,0 +1,269 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ + +import { isEqual, isPlainObject } from 'lodash-es' + +const SOURCE_NAMESPACE_LENGTH = 2 +const DEFINITION_INDENT = 2 + +const isNonBlankString = value => typeof value === 'string' && value.trim() !== '' + +const duplicatesOf = names => { + const seen = new Set() + const duplicates = new Set() + + names.forEach(name => { + if (seen.has(name)) { + duplicates.add(name) + } + seen.add(name) + }) + + return [...duplicates] +} + +const validateSource = (dataset, errors) => { + const { name, source } = dataset + + if (!isPlainObject(source)) { + errors.push(`Dataset "${name}" must have a source with a namespace and a name.`) + + return + } + + const { namespace } = source + + if (!Array.isArray(namespace) || namespace.length !== SOURCE_NAMESPACE_LENGTH || !namespace.every(isNonBlankString)) { + errors.push(`Dataset "${name}" must have a source namespace of exactly [catalog, schema].`) + } + + if (!isNonBlankString(source.name)) { + errors.push(`Dataset "${name}" must have a source with a non-empty name.`) + } +} + +const validateDatasets = (datasets, errors) => { + datasets.forEach((dataset, index) => { + if (!isPlainObject(dataset) || !isNonBlankString(dataset.name)) { + errors.push(`datasets[${index}] must have a non-empty name.`) + + return + } + + validateSource(dataset, errors) + + const fields = dataset.fields + if (fields !== undefined && !Array.isArray(fields)) { + errors.push(`Dataset "${dataset.name}" must declare fields as an array.`) + + return + } + + ;(fields || []).forEach((field, fieldIndex) => { + if (!isPlainObject(field) || !isNonBlankString(field.name)) { + errors.push(`datasets[${index}].fields[${fieldIndex}] must have a non-empty name.`) + } + }) + + const duplicateFields = duplicatesOf((fields || []).map(field => field?.name).filter(isNonBlankString)) + if (duplicateFields.length) { + errors.push(`Field names must be unique within dataset "${dataset.name}": ${duplicateFields.join(', ')}.`) + } + }) + + const duplicateDatasets = duplicatesOf(datasets.map(dataset => dataset?.name).filter(isNonBlankString)) + if (duplicateDatasets.length) { + errors.push(`Dataset names must be unique within the semantic model: ${duplicateDatasets.join(', ')}.`) + } +} + +const validateRelationships = (relationships, datasetNames, errors) => { + relationships.forEach((relationship, index) => { + if (!isPlainObject(relationship) || !isNonBlankString(relationship.name)) { + errors.push(`relationships[${index}] must have a non-empty name.`) + + return + } + + const { name, from, to, fromColumns, toColumns } = relationship + + ;[from, to].forEach(endpoint => { + if (!datasetNames.has(endpoint)) { + errors.push(`Relationship "${name}" references an unknown dataset "${endpoint}".`) + } + }) + + if (!Array.isArray(fromColumns) || !Array.isArray(toColumns) || !fromColumns.length || !toColumns.length) { + errors.push(`Relationship "${name}" must have non-empty fromColumns and toColumns.`) + + return + } + + if (fromColumns.length !== toColumns.length) { + errors.push(`Relationship "${name}" must have the same number of fromColumns and toColumns.`) + } + }) + + const duplicates = duplicatesOf(relationships.map(relationship => relationship?.name).filter(isNonBlankString)) + if (duplicates.length) { + errors.push(`Relationship names must be unique within the semantic model: ${duplicates.join(', ')}.`) + } +} + +const validateMetrics = (metrics, errors) => { + metrics.forEach((metric, index) => { + if (!isPlainObject(metric) || !isNonBlankString(metric.name)) { + errors.push(`metrics[${index}] must have a non-empty name.`) + } + }) + + const duplicates = duplicatesOf(metrics.map(metric => metric?.name).filter(isNonBlankString)) + if (duplicates.length) { + errors.push(`Metric names must be unique within the semantic model: ${duplicates.join(', ')}.`) + } +} + +/** + * Checks the structural rules the server also enforces, so obvious mistakes are reported before the + * request is sent. The server stays the authority: it additionally resolves every dataset source + * and its columns against the catalog, which the browser cannot do. + * + * @param {unknown} definition The parsed semantic model definition. + * @returns {string[]} The problems found, empty when the definition looks well formed. + */ +export const validateSemanticModelDefinition = definition => { + const errors = [] + + if (!isPlainObject(definition)) { + return ['The definition must be a JSON object.'] + } + + const { datasets, relationships, metrics } = definition + + if (!Array.isArray(datasets) || !datasets.length) { + return ['The definition must declare at least one dataset.'] + } + + validateDatasets(datasets, errors) + + const datasetNames = new Set(datasets.map(dataset => dataset?.name).filter(isNonBlankString)) + + if (relationships !== undefined) { + if (Array.isArray(relationships)) { + validateRelationships(relationships, datasetNames, errors) + } else { + errors.push('The definition must declare relationships as an array.') + } + } + + if (metrics !== undefined) { + if (Array.isArray(metrics)) { + validateMetrics(metrics, errors) + } else { + errors.push('The definition must declare metrics as an array.') + } + } + + return errors +} + +/** + * Parses the definition authored in the dialog and validates it. + * + * @param {string} text The raw JSON text. + * @returns {{ definition: object|null, errors: string[] }} The definition, or the problems found. + */ +export const parseSemanticModelDefinition = text => { + if (!isNonBlankString(text)) { + return { definition: null, errors: ['The definition is required.'] } + } + + let parsed + try { + parsed = JSON.parse(text) + } catch (err) { + return { definition: null, errors: [`The definition is not valid JSON: ${err.message}`] } + } + + const errors = validateSemanticModelDefinition(parsed) + + return errors.length ? { definition: null, errors } : { definition: parsed, errors: [] } +} + +/** + * Renders a definition for the JSON editor. + * + * @param {object|null|undefined} definition The definition to render. + * @returns {string} The pretty-printed definition, or an empty string when there is none. + */ +export const formatSemanticModelDefinition = definition => { + if (!definition) { + return '' + } + + return JSON.stringify(definition, null, DEFINITION_INDENT) +} + +/** + * Builds the atomic alter batch from what actually changed, so an unchanged definition is not + * revalidated against its sources. + * + * @param {object} originalData The loaded semantic model. + * @param {object} newData The edited semantic model. + * @returns {object[]} The ordered updates, empty when nothing changed. + */ +export const genSemanticModelUpdates = (originalData, newData) => { + const updates = [] + + if (originalData.name !== newData.name) { + updates.push({ '@type': 'rename', newName: newData.name }) + } + + const originalComment = originalData.comment || '' + const newComment = newData.comment || '' + + if (originalComment !== newComment) { + updates.push({ '@type': 'updateComment', newComment: newComment === '' ? null : newComment }) + } + + const originalProperties = originalData.properties || {} + const newProperties = newData.properties || {} + + for (const key in originalProperties) { + if (!(key in newProperties)) { Review Comment: [P2] Check only own keys when detecting removed properties `in` also checks `Object.prototype`. If a model has a valid property named `constructor` or `toString`, deleting its row still leaves that name `in {}`, so no `removeProperty` is generated. With no other edits, the dialog closes without a PUT and the property reappears when reopened; with other edits, it remains on the server. I reproduced this with the actual editor component. Use an own-property check such as `Object.hasOwn(newProperties, key)` and cover these names in the regression tests. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
