Merge pull request #152 from sat-utils/develop

publish 0.2.3

Author: matthewhanson

CHANGELOG.md

@@ -6,6 +6,16 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+## [v0.2.3] - 2019-01-29
+
+### Fixed
+- Proper handling of bounding box passed as string
+
+### Changed
+- De-normalize Item properties to include all properties from collection
+- Flattened elastic search to simplify query logic
+- Items returned will now include all 'Common' properties that are in the Items Collection
+
 ## [v0.2.2] - 2019-01-21
 
 ### Fixed
@@ -74,7 +84,8 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Refactor and improve splitting
 
 [Unreleased]: https://github.com/sat-utils/sat-api/compare/master...develop
-[v0.2.2]: https://github.com/sat-utils/sat-api/compare/v0.2.0...v0.2.2
+[v0.2.3]: https://github.com/sat-utils/sat-api/compare/v0.2.2...v0.2.3
+[v0.2.2]: https://github.com/sat-utils/sat-api/compare/v0.2.1...v0.2.2
 [v0.2.1]: https://github.com/sat-utils/sat-api/compare/v0.2.0...v0.2.1
 [v0.2.0]: https://github.com/sat-utils/sat-api/compare/v0.1.0...v0.2.0
 [v0.1.0]: https://github.com/sat-utils/sat-api/compare/v0.0.2...v0.1.0

lerna.json

@@ -1,6 +1,6 @@
 {
   "lerna": "2.11.0",
-  "version": "0.2.2",
+  "version": "0.2.3",
   "npmClient": "yarn",
   "packages": [
     "packages/*"

packages/api-lib/libs/api.js

@@ -38,7 +38,13 @@ const extractBbox = function (params) {
   let intersectsGeometry
   const { bbox } = params
   if (bbox) {
-    const boundingBox = extent(bbox)
+    let bboxArray
+    if (typeof bbox === 'string') {
+      bboxArray = JSON.parse(bbox)
+    } else {
+      bboxArray = bbox
+    }
+    const boundingBox = extent(bboxArray)
     const geojson = feature(boundingBox.polygon())
     intersectsGeometry = geojson
   }
@@ -236,15 +242,8 @@ const buildPageLinks = function (meta, parameters, endpoint) {
 }
 
 const searchItems = async function (parameters, page, limit, backend, endpoint) {
-  const arbitraryLimit = 5000
-  const { results: collectionResults } =
-    await backend.search(parameters, 'collections', 1, arbitraryLimit)
-  const collectionList = collectionResults.map((result) => result.id)
-  const collectionsQuery = Object.assign(
-    {}, parameters, { parentCollections: collectionList }
-  )
   const { results: itemsResults, meta: itemsMeta } =
-    await backend.search(collectionsQuery, 'items', page, limit)
+    await backend.search(parameters, 'items', page, limit)
   const pageLinks = buildPageLinks(itemsMeta, parameters, endpoint)
   const items = addItemLinks(itemsResults, endpoint)
   const response = wrapResponseInFeatureCollection(itemsMeta, items, pageLinks)

packages/api-lib/libs/es.js

@@ -8,7 +8,6 @@ const ElasticsearchWritableStream = require('./ElasticSearchWriteableStream')
 const logger = require('./logger')
 
 let _esClient
-
 /*
 This module is used for connecting to an Elasticsearch instance, writing records,
 searching records, and managing the indexes. It looks for the ES_HOST environment
@@ -77,7 +76,6 @@ async function esClient() {
 // Create STAC mappings
 async function prepare(index) {
   // TODO - different mappings for collection and item
-  let ready
   const props = {
     'type': 'object',
     properties: {
@@ -106,7 +104,6 @@ async function prepare(index) {
   }]
   const client = await esClient()
   const indexExists = await client.indices.exists({ index })
-
   if (!indexExists) {
     const payload = {
       index,
@@ -133,49 +130,71 @@ async function prepare(index) {
     try {
       await client.indices.create(payload)
       logger.info(`Created index: ${JSON.stringify(payload)}`)
-      ready = 0
     } catch (error) {
       const debugMessage = `Error creating index, already created: ${error}`
       logger.debug(debugMessage)
     }
   }
-  return ready
 }
 
 // Given an input stream and a transform, write records to an elasticsearch instance
 async function _stream() {
-  const toEs = through2.obj({ objectMode: true }, (data, encoding, next) => {
-    let index = ''
-    if (data && data.hasOwnProperty('extent')) {
-      index = 'collections'
-    } else if (data && data.hasOwnProperty('geometry')) {
-      index = 'items'
-    } else {
-      next()
-      return
-    }
-    // remove any hierarchy links in a non-mutating way
-    const hlinks = ['self', 'root', 'parent', 'child', 'collection', 'item']
-    const links = data.links.filter((link) => hlinks.includes(link))
-    const dataNoLinks = Object.assign({}, data, { links })
-
-    // create ES record
-    const record = {
-      index,
-      type: 'doc',
-      id: dataNoLinks.id,
-      action: 'update',
-      _retry_on_conflict: 3,
-      body: {
-        doc: dataNoLinks,
-        doc_as_upsert: true
-      }
-    }
-    next(null, record)
-  })
   let esStreams
   try {
+    let collections = []
     const client = await esClient()
+    const indexExists = await client.indices.exists({ index: 'collections' })
+    if (indexExists) {
+      const body = { query: { match_all: {} } }
+      const searchParams = {
+        index: 'collections',
+        body
+      }
+      const resultBody = await client.search(searchParams)
+      collections = resultBody.hits.hits.map((r) => (r._source))
+    }
+
+    const toEs = through2.obj({ objectMode: true }, (data, encoding, next) => {
+      let index = ''
+      if (data && data.hasOwnProperty('extent')) {
+        index = 'collections'
+      } else if (data && data.hasOwnProperty('geometry')) {
+        index = 'items'
+      } else {
+        next()
+        return
+      }
+      // remove any hierarchy links in a non-mutating way
+      const hlinks = ['self', 'root', 'parent', 'child', 'collection', 'item']
+      const links = data.links.filter((link) => hlinks.includes(link))
+      let esDataObject = Object.assign({}, data, { links })
+      if (index === 'items') {
+        const collectionId = data.properties.collection
+        const itemCollection =
+          collections.find((collection) => (collectionId === collection.id))
+        if (itemCollection) {
+          const flatProperties =
+            Object.assign({}, itemCollection.properties, data.properties)
+          esDataObject = Object.assign({}, esDataObject, { properties: flatProperties })
+        } else {
+          logger.error(`${data.id} has no collection`)
+        }
+      }
+
+      // create ES record
+      const record = {
+        index,
+        type: 'doc',
+        id: esDataObject.id,
+        action: 'update',
+        _retry_on_conflict: 3,
+        body: {
+          doc: esDataObject,
+          doc_as_upsert: true
+        }
+      }
+      next(null, record)
+    })
     const esStream = new ElasticsearchWritableStream({ client: client }, {
       objectMode: true,
       highWaterMark: process.env.ES_BATCH_SIZE || 500
@@ -243,9 +262,11 @@ function buildDatetimeQuery(parameters) {
 
 function buildQuery(parameters) {
   const eq = 'eq'
-  const { query, parentCollections, intersects } = parameters
+  const { query, intersects } = parameters
   let must = []
   if (query) {
+    // Using reduce rather than map as we don't currently support all
+    // stac query operators.
     must = Object.keys(query).reduce((accumulator, property) => {
       const operatorsObject = query[property]
       const operators = Object.keys(operatorsObject)
@@ -265,6 +286,7 @@ function buildQuery(parameters) {
       return accumulator
     }, must)
   }
+
   if (intersects) {
     const { geometry } = intersects
     must.push({
@@ -279,19 +301,7 @@ function buildQuery(parameters) {
     must.push(datetimeQuery)
   }
 
-  let filter
-  if (parentCollections && parentCollections.length !== 0) {
-    filter = {
-      bool: {
-        should: [
-          { terms: { 'properties.collection': parentCollections } },
-          { bool: { must } }
-        ]
-      }
-    }
-  } else {
-    filter = { bool: { must } }
-  }
+  const filter = { bool: { must } }
   const queryBody = {
     constant_score: { filter }
   }

packages/api-lib/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sat-utils/api-lib",
-  "version": "0.2.2",
+  "version": "0.2.3",
   "description": "A library for creating a search API of public Satellites metadata using Elasticsearch",
   "main": "index.js",
   "scripts": {

packages/api-lib/tests/fixtures/stac/collection2.json

@@ -220,7 +220,7 @@
         "eo:gsd": 15,
         "eo:instrument": "OLI_TIRS",
         "eo:off_nadir": 0,
-        "eo:platform": "landsat-8"
+        "eo:platform": "platform2"
     },
     "providers": [
         {

packages/api-lib/tests/integration/ingestCollections.js

@@ -0,0 +1,9 @@
+process.env.ES_HOST = `http://${process.env.DOCKER_NAME}:4571`
+const ingest = require('../../libs/ingest').ingest
+const backend = require('../../libs/es')
+
+async function doIngest() {
+  await ingest('../fixtures/stac/catalog.json', backend, true, true)
+  console.log('Collections done')
+}
+doIngest()

packages/api-lib/tests/integration/ingestData.js

@@ -2,6 +2,9 @@ process.env.ES_HOST = `http://${process.env.DOCKER_NAME}:4571`
 const ingest = require('../../libs/ingest').ingest
 const backend = require('../../libs/es')
 
-//ingest('../fixtures/stac/catalog.json', backend)
-ingest('https://landsat-stac.s3.amazonaws.com/landsat-8-l1/catalog.json', backend)
-
+async function doIngest() {
+  await ingest('../fixtures/stac/catalog.json', backend)
+  console.log('Items done')
+}
+//ingest('https://landsat-stac.s3.amazonaws.com/landsat-8-l1/catalog.json', backend)
+doIngest()

packages/api-lib/tests/integration/runIntegration.sh

@@ -1,4 +1,4 @@
 #!/bin/bash
 docker-compose up & while ! nc -z $DOCKER_NAME 4571; do sleep 1; done;
 sleep 20;
-node ./ingestData.js && yarn ava ./tests/integration/test_api.js
+node ./ingestCollections.js && node ./ingestData.js && yarn ava ./tests/integration/test_api.js

packages/api-lib/tests/integration/test_api.js

@@ -151,3 +151,27 @@ test('stac/search sort', async (t) => {
   }, backend, endpoint)
   t.is(response.features[0].id, 'LC80100102015082LGN00')
 })
+
+test('stac/search flattened collection properties', async (t) => {
+  let response = await search('/stac/search', {
+    query: {
+      'eo:platform': {
+        eq: 'platform2'
+      }
+    }
+  }, backend, endpoint)
+  t.is(response.features[0].id, 'collection2_item')
+
+  response = await search('/stac/search', {
+    query: {
+      'eo:platform': {
+        eq: 'landsat-8'
+      }
+    }
+  }, backend, endpoint)
+  const havePlatform =
+    response.features.filter(
+      (item) => (item.properties['eo:platform'] === 'landsat-8')
+    )
+  t.is(havePlatform.length, response.features.length)
+})