@abi-software/map-side-bar 2.10.7-demo.1 → 2.11.2-acupoints.0

package/src/services/flatmapQueries.js

@@ -0,0 +1,498 @@
+/* eslint-disable no-alert, no-console */
+// remove duplicates by stringifying the objects
+const removeDuplicates = function (arrayOfAnything) {
+  if (!arrayOfAnything) return []
+  return [...new Set(arrayOfAnything.map((e) => JSON.stringify(e)))].map((e) =>
+    JSON.parse(e)
+  )
+}
+
+const cachedLabels = {}
+const cachedTaxonLabels = [];
+
+const findTaxonomyLabel = async function (flatmapAPI, taxonomy) {
+  if (cachedLabels && cachedLabels.hasOwnProperty(taxonomy)) {
+    return cachedLabels[taxonomy]
+  }
+
+  return new Promise((resolve) => {
+    fetch(`${flatmapAPI}knowledge/label/${taxonomy}`, {
+      method: 'GET',
+    })
+      .then((response) => response.json())
+      .then((data) => {
+        let label = data.label
+        if (label === 'Mammalia') {
+          label = 'Mammalia not otherwise specified'
+        }
+        cachedLabels[taxonomy] = label
+        resolve(label)
+      })
+      .catch((error) => {
+        console.error('Error:', error)
+        cachedLabels[taxonomy] = taxonomy
+        resolve(taxonomy)
+      })
+  })
+}
+
+const findTaxonomyLabels = async function (mapImp, taxonomies) {
+  const intersectionTaxonomies = taxonomies.filter((taxonomy) =>
+    cachedTaxonLabels.some((obj) => obj.taxon === taxonomy)
+  );
+
+  const foundCachedTaxonLabels = cachedTaxonLabels.filter((obj) =>
+    intersectionTaxonomies.includes(obj.taxon)
+  );
+
+  const leftoverTaxonomies = taxonomies.filter((taxonomy) =>
+    !intersectionTaxonomies.includes(taxonomy)
+  );
+
+  if (!leftoverTaxonomies.length) {
+    return foundCachedTaxonLabels;
+  } else {
+    const entityLabels = await mapImp.queryLabels(leftoverTaxonomies);
+    if (entityLabels.length) {
+      entityLabels.forEach((entityLabel) => {
+        let { entity: taxon, label } = entityLabel;
+        if (label === 'Mammalia') {
+          label = 'Mammalia not otherwise specified'
+        }
+        const item = { taxon, label };
+        foundCachedTaxonLabels.push(item);
+        cachedTaxonLabels.push(item);
+      });
+      return foundCachedTaxonLabels;
+    }
+  }
+}
+
+const inArray = function (ar1, ar2) {
+  if (!ar1 || !ar2) return false
+  let as1 = JSON.stringify(ar1)
+  let as2 = JSON.stringify(ar2)
+  return as1.indexOf(as2) !== -1
+}
+
+let FlatmapQueries = function () {
+  this.initialise = function (flatmapApi) {
+    this.flatmapApi = flatmapApi
+    this.destinations = []
+    this.origins = []
+    this.components = []
+    this.rawURLs = []
+    this.controller = undefined
+    this.uberons = []
+    this.lookUp = []
+  }
+
+  this.createTooltipData = async function (mapImp, eventData) {
+    let hyperlinks = []
+    if (
+      eventData.feature.hyperlinks &&
+      eventData.feature.hyperlinks.length > 0
+    ) {
+      hyperlinks = eventData.feature.hyperlinks
+    } else {
+      hyperlinks = this.rawURLs;
+    }
+    let taxonomyLabel = undefined
+    if (eventData.provenanceTaxonomy) {
+      taxonomyLabel = []
+      const entityLabels = await findTaxonomyLabels(mapImp, eventData.provenanceTaxonomy);
+      if (entityLabels.length) {
+        entityLabels.forEach((entityLabel) => {
+          const { label } = entityLabel;
+          taxonomyLabel.push(label);
+        });
+      }
+    }
+
+    let tooltipData = {
+      destinations: this.destinations,
+      origins: this.origins,
+      components: this.components,
+      destinationsWithDatasets: this.destinationsWithDatasets,
+      originsWithDatasets: this.originsWithDatasets,
+      componentsWithDatasets: this.componentsWithDatasets,
+      title: eventData.label,
+      featureId: eventData.resource,
+      hyperlinks: hyperlinks,
+      provenanceTaxonomy: eventData.provenanceTaxonomy,
+      provenanceTaxonomyLabel: taxonomyLabel,
+    }
+    return tooltipData
+  }
+
+  this.createComponentsLabelList = function (components, lookUp) {
+    let labelList = []
+    components.forEach((n) => {
+      labelList.push(this.createLabelFromNeuralNode(n[0]), lookUp)
+      if (n.length === 2) {
+        labelList.push(this.createLabelFromNeuralNode(n[1]), lookUp)
+      }
+    })
+    return labelList
+  }
+
+  this.createLabelLookup = function (mapImp, uberons) {
+    return new Promise(async (resolve) => {
+      let uberonMap = {}
+      this.uberons = []
+      const entityLabels = await findTaxonomyLabels(mapImp, uberons);
+      if (entityLabels.length) {
+        entityLabels.forEach((entityLabel) => {
+          const { taxon: entity, label } = entityLabel;
+          uberonMap[entity] = label;
+          this.uberons.push({
+            id: entity,
+            name: label,
+          })
+        });
+        resolve(uberonMap)
+      }
+    })
+  }
+
+  this.buildConnectivitySqlStatement = function (keastIds) {
+    let sql = 'select knowledge from knowledge where entity in ('
+    if (keastIds.length === 1) {
+      sql += `'${keastIds[0]}')`
+    } else if (keastIds.length > 1) {
+      for (let i in keastIds) {
+        sql += `'${keastIds[i]}'${i >= keastIds.length - 1 ? ')' : ','} `
+      }
+    }
+    return sql
+  }
+
+  this.buildLabelSqlStatement = function (uberons) {
+    let sql = 'select entity, label from labels where entity in ('
+    if (uberons.length === 1) {
+      sql += `'${uberons[0]}')`
+    } else if (uberons.length > 1) {
+      for (let i in uberons) {
+        sql += `'${uberons[i]}'${i >= uberons.length - 1 ? ')' : ','} `
+      }
+    }
+    return sql
+  }
+
+  this.findAllIdsFromConnectivity = function (connectivity) {
+    let dnodes = connectivity.connectivity.flat() // get nodes from edgelist
+    let nodes = [...new Set(dnodes)] // remove duplicates
+    let found = []
+    nodes.forEach((n) => {
+      if (Array.isArray(n)) {
+        found.push(n.flat())
+      } else {
+        found.push(n)
+      }
+    })
+    return [...new Set(found.flat())]
+  }
+
+  this.flattenConntectivity = function (connectivity) {
+    let dnodes = connectivity.flat() // get nodes from edgelist
+    let nodes = [...new Set(dnodes)] // remove duplicates
+    let found = []
+    nodes.forEach((n) => {
+      if (Array.isArray(n)) {
+        found.push(n.flat())
+      } else {
+        found.push(n)
+      }
+    })
+    return found.flat()
+  }
+
+  this.findComponents = function (connectivity) {
+    let dnodes = connectivity.connectivity.flat() // get nodes from edgelist
+    let nodes = removeDuplicates(dnodes)
+
+    let found = []
+    let terminal = false
+    nodes.forEach((node) => {
+      terminal = false
+      // Check if the node is an destination or origin (note that they are labelled dendrite and axon as opposed to origin and destination)
+      if (inArray(connectivity.axons, node)) {
+        terminal = true
+      }
+      if (connectivity.somas && inArray(connectivity.somas, node)) {
+        terminal = true
+      }
+      if (inArray(connectivity.dendrites, node)) {
+        terminal = true
+      }
+      if (!terminal) {
+        found.push(node)
+      }
+    })
+
+    return found
+  }
+
+  this.retrieveFlatmapKnowledgeForEvent = async function (mapImp, eventData) {
+    // check if there is an existing query
+    if (this.controller) this.controller.abort()
+
+    // set up the abort controller
+    this.controller = new AbortController()
+    const signal = this.controller.signal
+
+    const keastIds = eventData.resource
+    this.destinations = []
+    this.origins = []
+    this.components = []
+    this.rawURLs = []
+    if (!keastIds || keastIds.length == 0 || !keastIds[0]) return
+
+    let prom1 = this.queryForConnectivityNew(mapImp, keastIds, signal) // This on returns a promise so dont need 'await'
+    let results = await Promise.all([prom1])
+    return results
+  }
+
+  this.queryForConnectivityNew = function (mapImp, keastIds, signal, processConnectivity=true) {
+    return new Promise((resolve) => {
+      mapImp.queryKnowledge(keastIds[0])
+        .then((response) => {
+          if (this.checkConnectivityExists(response)) {
+            let connectivity = response;
+            if (processConnectivity) {
+              this.processConnectivity(mapImp, connectivity).then((processedConnectivity) => {
+                // response.references is publication urls
+                if (response.references) {
+                  // with publications from both PubMed and Others
+                  this.rawURLs = [...response.references];
+                  resolve(processedConnectivity)
+                } else {
+                  // without publications
+                  resolve(processedConnectivity)
+                }
+              })
+            }
+            else resolve(connectivity)
+          } else {
+            resolve(false)
+          }
+        })
+        .catch((error) => {
+          if (error.name === 'AbortError') {
+            // This error is from AbortController's abort method.
+          } else {
+            // console.error('Error:', error)
+            // TODO: to update after queryKnowledge method update
+            console.warn(`Unable to get the knowledge for the entity ${keastIds[0]}.`)
+          }
+          resolve(false)
+        })
+    })
+  }
+
+  this.queryForConnectivity = function (mapImp, keastIds, signal, processConnectivity=true) {
+    const data = { sql: this.buildConnectivitySqlStatement(keastIds) }
+    const headers = {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+      },
+      body: JSON.stringify(data),
+      ...(signal ? { signal: signal } : {}), // add signal to header if it exists
+    }
+    return new Promise((resolve) => {
+      fetch(`${this.flatmapApi}knowledge/query/`, headers)
+        .then((response) => response.json())
+        .then((data) => {
+          if (this.connectivityExists(data)) {
+            let connectivity = JSON.parse(data.values[0][0])
+            if (processConnectivity) {
+              this.processConnectivity(mapImp, connectivity).then((processedConnectivity) => {
+                resolve(processedConnectivity)
+              })
+            }
+            else resolve(connectivity)
+          } else {
+            resolve(false)
+          }
+        })
+        .catch((error) => {
+          if (error.name === 'AbortError') {
+            // This error is from AbortController's abort method.
+          } else {
+            console.error('Error:', error)
+          }
+          resolve(false)
+        })
+    })
+  }
+
+  this.checkConnectivityExists = function (data) {
+    return data && data.connectivity?.length;
+  };
+
+  this.connectivityExists = function (data) {
+    if (
+      data.values &&
+      data.values.length > 0 &&
+      JSON.parse(data.values[0][0]).connectivity &&
+      JSON.parse(data.values[0][0]).connectivity.length > 0
+    ) {
+      return true
+    } else {
+      return false
+    }
+  }
+
+  // This function is used to determine if a node is a single node or a node with multiple children
+  //  Returns the id of the node if it is a single node, otherwise returns false
+  this.findIfNodeIsSingle = function (node) {
+    if (node.length === 1) { // If the node is in the form [id]
+      console.error("Server returns a single node", node)
+      return node[0]
+    } else {
+      if (node.length === 2 && node[1].length === 0) { // If the node is in the form [id, []]
+        return node[0]
+      } else {
+        return false // If the node is in the form [id, [id1, id2]]
+      }
+    }
+  }
+
+  this.createLabelFromNeuralNode = function (node, lookUp) {
+
+    // Check if the node is a single node or a node with multiple children
+    let nodeIsSingle = this.findIfNodeIsSingle(node)
+
+    // Case where node is in the form [id]
+    if (nodeIsSingle) {
+      return lookUp[nodeIsSingle]
+    }
+
+    // Case where node is in the form [id, [id1 (,id2)]]
+    let label = lookUp[node[0]]
+    if (node.length === 2 && node[1].length > 0) {
+      node[1].forEach((n) => {
+        if (lookUp[n] == undefined) {
+          label += `, ${n}`
+        } else {
+          label += `, ${lookUp[n]}`
+        }
+      })
+    }
+    return label
+  }
+
+  this.flattenAndFindDatasets = function (components, axons, dendrites) {
+    // process the nodes for finding datasets (Note this is not critical to the tooltip, only for the 'search on components' button)
+    let componentsFlat = this.flattenConntectivity(components)
+    let axonsFlat = this.flattenConntectivity(axons)
+    let dendritesFlat = this.flattenConntectivity(dendrites)
+
+    // Filter for the anatomy which is annotated on datasets
+    this.destinationsWithDatasets = this.uberons.filter(
+      (ub) => axonsFlat.indexOf(ub.id) !== -1
+    )
+    this.originsWithDatasets = this.uberons.filter(
+      (ub) => dendritesFlat.indexOf(ub.id) !== -1
+    )
+    this.componentsWithDatasets = this.uberons.filter(
+      (ub) => componentsFlat.indexOf(ub.id) !== -1
+    )
+  }
+
+  this.processConnectivity = function (mapImp, connectivity) {
+    return new Promise((resolve) => {
+      // Filter the origin and destinations from components
+      let components = this.findComponents(connectivity)
+
+      // Remove duplicates
+      let axons = removeDuplicates(connectivity.axons)
+      //Somas will become part of origins, support this for future proof
+      let dendrites = []
+      if (connectivity.somas && connectivity.somas.length > 0) {
+        dendrites.push(...connectivity.somas)
+      }
+      if (connectivity.dendrites && connectivity.dendrites.length > 0) {
+        dendrites.push(...connectivity.dendrites)
+      }
+      dendrites = removeDuplicates(dendrites)
+
+      // Create list of ids to get labels for
+      let conIds = this.findAllIdsFromConnectivity(connectivity)
+
+      // Create readable labels from the nodes. Setting this to 'this.origins' updates the display
+      this.createLabelLookup(mapImp, conIds).then((lookUp) => {
+        this.destinations = axons.map((a) =>
+          this.createLabelFromNeuralNode(a, lookUp)
+        )
+        this.origins = dendrites.map((d) =>
+          this.createLabelFromNeuralNode(d, lookUp)
+        )
+        this.components = components.map((c) =>
+          this.createLabelFromNeuralNode(c, lookUp)
+        )
+        this.flattenAndFindDatasets(components, axons, dendrites)
+        resolve({
+          ids: {
+            axons: axons,
+            dendrites: dendrites,
+            components: components,
+          },
+          labels: {
+            destinations: this.destinations,
+            origins: this.origins,
+            components: this.components,
+          }
+        })
+      })
+    })
+  }
+
+  this.flattenConntectivity = function (connectivity) {
+    let dnodes = connectivity.flat() // get nodes from edgelist
+    let nodes = [...new Set(dnodes)] // remove duplicates
+    let found = []
+    nodes.forEach((n) => {
+      if (Array.isArray(n)) {
+        found.push(n.flat())
+      } else {
+        found.push(n)
+      }
+    })
+    return found.flat()
+  }
+
+  this.buildPubmedSqlStatement = function (keastIds) {
+    let sql = 'select distinct publication from publications where entity in ('
+    if (keastIds.length === 1) {
+      sql += `'${keastIds[0]}')`
+    } else if (keastIds.length > 1) {
+      for (let i in keastIds) {
+        sql += `'${keastIds[i]}'${i >= keastIds.length - 1 ? ')' : ','} `
+      }
+    }
+    return sql
+  }
+
+  this.buildPubmedSqlStatementForModels = function (model) {
+    return `select distinct publication from publications where entity = '${model}'`
+  }
+
+  this.flatmapQuery = function (sql) {
+    const data = { sql: sql }
+    return fetch(`${this.flatmapApi}knowledge/query/`, {
+      method: 'POST',
+      headers: {
+        'Content-Type': 'application/json',
+      },
+      body: JSON.stringify(data),
+    })
+      .then((response) => response.json())
+      .catch((error) => {
+        console.error('Error:', error)
+      })
+  }
+}
+
+export { FlatmapQueries, findTaxonomyLabel, findTaxonomyLabels }