feat(#373): merge-contacts action (#647)

This adds new action cht merge-contacts. The change proposes to:

Move the code for move-contacts into a library lib/hierarchy-operations with interfaces move and merge
Parameterize the move-contacts code changing logic for: input validation, deletes the top-level contact, changes how lineages are updated, and adds report reassignment. All else remains unaltered.
Handles reports and contacts only. Unclear what other doc types I should be worried about.

.gitignore

@@ -5,6 +5,7 @@ upload-docs.*.log.json
 /.vscode/
 /.idea/
 /.settings/
+/json_docs/
 *.swp
 coverage
 .nyc_output

src/fn/merge-contacts.js

@@ -0,0 +1,69 @@
+const minimist = require('minimist');
+const path = require('path');
+
+const environment = require('../lib/environment');
+const pouch = require('../lib/db');
+const { info } = require('../lib/log');
+
+const HierarchyOperations = require('../lib/hierarchy-operations');
+
+module.exports = {
+  requiresInstance: true,
+  execute: () => {
+    const args = parseExtraArgs(environment.pathToProject, environment.extraArgs);
+    const db = pouch();
+    const options = {
+      docDirectoryPath: args.docDirectoryPath,
+      force: args.force,
+    };
+    return HierarchyOperations(db, options).merge(args.sourceIds, args.destinationId);
+  }
+};
+
+// Parses extraArgs and asserts if required parameters are not present
+const parseExtraArgs = (projectDir, extraArgs = []) => {
+  const args = minimist(extraArgs, { boolean: true });
+
+  const sourceIds = (args.sources || args.source || '')
+    .split(',')
+    .filter(Boolean);
+
+  if (!args.destination) {
+    usage();
+    throw Error(`Action "merge-contacts" is missing required contact ID ${bold('--destination')}. Other contacts will be merged into this contact.`);
+  }
+
+  if (sourceIds.length === 0) {
+    usage();
+    throw Error(`Action "merge-contacts" is missing required contact ID(s) ${bold('--sources')}. These contacts will be merged into the contact specified by ${bold('--destination')}`);
+  }
+
+  return {
+    destinationId: args.destination,
+    sourceIds,
+    docDirectoryPath: path.resolve(projectDir, args.docDirectoryPath || 'json_docs'),
+    force: !!args.force,
+  };
+};
+
+const bold = text => `\x1b[1m${text}\x1b[0m`;
+const usage = () => {
+  info(`
+${bold('cht-conf\'s merge-contacts action')}
+When combined with 'upload-docs' this action moves all of the contacts and reports under ${bold('sources')} to be under ${bold('destination')}.
+The top-level ${bold('source contact(s)')} are deleted and no data from this document is merged or preserved.
+
+${bold('USAGE')}
+cht --local merge-contacts -- --destination=<destination_id> --sources=<source_id1>,<source_id2>
+
+${bold('OPTIONS')}
+--destination=<destination_id>
+  Specifies the ID of the contact that should receive the moving contacts and reports.
+
+--sources=<source_id1>,<source_id2>
+  A comma delimited list of IDs of contacts which will be deleted. The hierarchy of contacts and reports under it will be moved to be under the destination contact.
+
+--docDirectoryPath=<path to stage docs>
+  Specifies the folder used to store the documents representing the changes in hierarchy.
+`);
+};

src/fn/move-contacts.js

@@ -1,105 +1,34 @@
 const minimist = require('minimist');
 const path = require('path');
-const userPrompt = require('../lib/user-prompt');
 
 const environment = require('../lib/environment');
-const fs = require('../lib/sync-fs');
-const lineageManipulation = require('../lib/lineage-manipulation');
-const lineageConstraints = require('../lib/lineage-constraints');
 const pouch = require('../lib/db');
-const { warn, trace, info } = require('../lib/log');
+const { info } = require('../lib/log');
 
-const HIERARCHY_ROOT = 'root';
-const BATCH_SIZE = 10000;
+const HierarchyOperations = require('../lib/hierarchy-operations');
 
 module.exports = {
   requiresInstance: true,
   execute: () => {
     const args = parseExtraArgs(environment.pathToProject, environment.extraArgs);
     const db = pouch();
-    prepareDocumentDirectory(args);
-    return updateLineagesAndStage(args, db);
+    const options = {
+      docDirectoryPath: args.docDirectoryPath,
+      force: args.force,
+    };
+    return HierarchyOperations(db, options).move(args.sourceIds, args.destinationId);
   }
 };
 
-const prettyPrintDocument = doc => `'${doc.name}' (${doc._id})`;
-const updateLineagesAndStage = async (options, db) => {
-  trace(`Fetching contact details for parent: ${options.parentId}`);
-  const parentDoc = await fetch.contact(db, options.parentId);
-
-  const constraints = await lineageConstraints(db, parentDoc);
-  const contactDocs = await fetch.contactList(db, options.contactIds);
-  await validateContacts(contactDocs, constraints);
-
-  let affectedContactCount = 0, affectedReportCount = 0;
-  const replacementLineage = lineageManipulation.createLineageFromDoc(parentDoc);
-  for (let contactId of options.contactIds) {
-    const contactDoc = contactDocs[contactId];
-    const descendantsAndSelf = await fetch.descendantsOf(db, contactId);
-
-    // Check that primary contact is not removed from areas where they are required
-    const invalidPrimaryContactDoc = await constraints.getPrimaryContactViolations(contactDoc, descendantsAndSelf);
-    if (invalidPrimaryContactDoc) {
-      throw Error(`Cannot remove contact ${prettyPrintDocument(invalidPrimaryContactDoc)} from the hierarchy for which they are a primary contact.`);
-    }
-
-    trace(`Considering lineage updates to ${descendantsAndSelf.length} descendant(s) of contact ${prettyPrintDocument(contactDoc)}.`);
-    const updatedDescendants = replaceLineageInContacts(descendantsAndSelf, replacementLineage, contactId);
-
-    const ancestors = await fetch.ancestorsOf(db, contactDoc);
-    trace(`Considering primary contact updates to ${ancestors.length} ancestor(s) of contact ${prettyPrintDocument(contactDoc)}.`);
-    const updatedAncestors = replaceLineageInAncestors(descendantsAndSelf, ancestors);
-
-    minifyLineageAndWriteToDisk([...updatedDescendants, ...updatedAncestors], options);
-
-    const movedReportsCount = await moveReports(db, descendantsAndSelf, options, replacementLineage, contactId);
-    trace(`${movedReportsCount} report(s) created by these affected contact(s) will be updated`);
-
-    affectedContactCount += updatedDescendants.length + updatedAncestors.length;
-    affectedReportCount += movedReportsCount;
-
-    info(`Staged updates to ${prettyPrintDocument(contactDoc)}. ${updatedDescendants.length} contact(s) and ${movedReportsCount} report(s).`);
-  }
-
-  info(`Staged changes to lineage information for ${affectedContactCount} contact(s) and ${affectedReportCount} report(s).`);
-};
-
-/*
-Checks for any errors which this will create in the hierarchy (hierarchy schema, circular hierarchies)
-Confirms the list of contacts are possible to move
-*/
-const validateContacts = async (contactDocs, constraints) => {
-  Object.values(contactDocs).forEach(doc => {
-    const hierarchyError = constraints.getHierarchyErrors(doc);
-    if (hierarchyError) {
-      throw Error(`Hierarchy Constraints: ${hierarchyError}`);
-    }
-  });
-
-  /*
-  It is nice that the tool can move lists of contacts as one operation, but strange things happen when two contactIds are in the same lineage.
-  For example, moving a district_hospital and moving a contact under that district_hospital to a new clinic causes multiple colliding writes to the same json file.
-  */
-  const contactIds = Object.keys(contactDocs);
-  Object.values(contactDocs)
-    .forEach(doc => {
-      const parentIdsOfDoc = (doc.parent && lineageManipulation.pluckIdsFromLineage(doc.parent)) || [];
-      const violatingParentId = parentIdsOfDoc.find(parentId => contactIds.includes(parentId));
-      if (violatingParentId) {
-        throw Error(`Unable to move two documents from the same lineage: '${doc._id}' and '${violatingParentId}'`);
-      }
-    });
-};
-
 // Parses extraArgs and asserts if required parameters are not present
 const parseExtraArgs = (projectDir, extraArgs = []) => {
   const args = minimist(extraArgs, { boolean: true });
 
-  const contactIds = (args.contacts || args.contact || '')
+  const sourceIds = (args.contacts || args.contact || '')
     .split(',')
     .filter(id => id);
 
-  if (contactIds.length === 0) {
+  if (sourceIds.length === 0) {
     usage();
     throw Error('Action "move-contacts" is missing required list of contacts to be moved');
   }
@@ -110,28 +39,15 @@ const parseExtraArgs = (projectDir, extraArgs = []) => {
   }
 
   return {
-    parentId: args.parent,
-    contactIds,
+    destinationId: args.parent,
+    sourceIds,
     docDirectoryPath: path.resolve(projectDir, args.docDirectoryPath || 'json_docs'),
     force: !!args.force,
   };
 };
 
-const prepareDocumentDirectory = ({ docDirectoryPath, force }) => {
-  if (!fs.exists(docDirectoryPath)) {
-    fs.mkdir(docDirectoryPath);
-  } else if (!force && fs.recurseFiles(docDirectoryPath).length > 0) {
-    warn(`The document folder '${docDirectoryPath}' already contains files. It is recommended you start with a clean folder. Do you want to delete the contents of this folder and continue?`);
-    if(userPrompt.keyInYN()) {
-      fs.deleteFilesInFolder(docDirectoryPath);
-    } else {
-      throw new Error('User aborted execution.');
-    }
-  }
-};
-
+const bold = text => `\x1b[1m${text}\x1b[0m`;
 const usage = () => {
-  const bold = text => `\x1b[1m${text}\x1b[0m`;
   info(`
 ${bold('cht-conf\'s move-contacts action')}
 When combined with 'upload-docs' this action effectively moves a contact from one place in the hierarchy to another.
@@ -144,148 +60,9 @@ ${bold('OPTIONS')}
   A comma delimited list of ids of contacts to be moved.
 
 --parent=<parent_id>
-  Specifies the ID of the new parent. Use '${HIERARCHY_ROOT}' to identify the top of the hierarchy (no parent).
+  Specifies the ID of the new parent. Use '${HierarchyOperations.HIERARCHY_ROOT}' to identify the top of the hierarchy (no parent).
 
 --docDirectoryPath=<path to stage docs>
   Specifies the folder used to store the documents representing the changes in hierarchy.
 `);
 };
-
-const moveReports = async (db, descendantsAndSelf, writeOptions, replacementLineage, contactId) => {
-  const contactIds = descendantsAndSelf.map(contact => contact._id);
-
-  let skip = 0;
-  let reportDocsBatch;
-  do {
-    info(`Processing ${skip} to ${skip + BATCH_SIZE} report docs`);
-    reportDocsBatch = await fetch.reportsCreatedBy(db, contactIds, skip);
-
-    const updatedReports = replaceLineageInReports(reportDocsBatch, replacementLineage, contactId);
-    minifyLineageAndWriteToDisk(updatedReports, writeOptions);
-
-    skip += reportDocsBatch.length;
-  } while (reportDocsBatch.length >= BATCH_SIZE);
-
-  return skip;
-};
-
-const minifyLineageAndWriteToDisk = (docs, parsedArgs) => {
-  docs.forEach(doc => {
-    lineageManipulation.minifyLineagesInDoc(doc);
-    writeDocumentToDisk(parsedArgs, doc);
-  });
-};
-
-const writeDocumentToDisk = ({ docDirectoryPath }, doc) => {
-  const destinationPath = path.join(docDirectoryPath, `${doc._id}.doc.json`);
-  if (fs.exists(destinationPath)) {
-    warn(`File at ${destinationPath} already exists and is being overwritten.`);
-  }
-
-  trace(`Writing updated document to ${destinationPath}`);
-  fs.writeJson(destinationPath, doc);
-};
-
-const fetch = {
-  /*
-  Fetches all of the documents associated with the "contactIds" and confirms they exist.
-  */
-  contactList: async (db, ids) => {
-    const contactDocs = await db.allDocs({
-      keys: ids,
-      include_docs: true,
-    });
-
-    const missingContactErrors = contactDocs.rows.filter(row => !row.doc).map(row => `Contact with id '${row.key}' could not be found.`);
-    if (missingContactErrors.length > 0) {
-      throw Error(missingContactErrors);
-    }
-
-    return contactDocs.rows.reduce((agg, curr) => Object.assign(agg, { [curr.doc._id]: curr.doc }), {});
-  },
-
-  contact: async (db, id) => {
-    try {
-      if (id === HIERARCHY_ROOT) {
-        return undefined;
-      }
-
-      return await db.get(id);
-    } catch (err) {
-      if (err.name !== 'not_found') {
-        throw err;
-      }
-
-      throw Error(`Contact with id '${id}' could not be found`);
-    }
-  },
-
-  /*
-  Given a contact's id, obtain the documents of all descendant contacts
-  */
-  descendantsOf: async (db, contactId) => {
-    const descendantDocs = await db.query('medic/contacts_by_depth', {
-      key: [contactId],
-      include_docs: true,
-    });
-
-    return descendantDocs.rows
-      .map(row => row.doc)
-      /* We should not move or update tombstone documents */
-      .filter(doc => doc && doc.type !== 'tombstone');
-  },
-
-  reportsCreatedBy: async (db, contactIds, skip) => {
-    const reports = await db.query('medic-client/reports_by_freetext', {
-      keys: contactIds.map(id => [`contact:${id}`]),
-      include_docs: true,
-      limit: BATCH_SIZE,
-      skip: skip,
-    });
-
-    return reports.rows.map(row => row.doc);
-  },
-
-  ancestorsOf: async (db, contactDoc) => {
-    const ancestorIds = lineageManipulation.pluckIdsFromLineage(contactDoc.parent);
-    const ancestors = await db.allDocs({
-      keys: ancestorIds,
-      include_docs: true,
-    });
-
-    const ancestorIdsNotFound = ancestors.rows.filter(ancestor => !ancestor.doc).map(ancestor => ancestor.key);
-    if (ancestorIdsNotFound.length > 0) {
-      throw Error(`Contact '${prettyPrintDocument(contactDoc)} has parent id(s) '${ancestorIdsNotFound.join(',')}' which could not be found.`);
-    }
-
-    return ancestors.rows.map(ancestor => ancestor.doc);
-  },
-};
-
-const replaceLineageInReports = (reportsCreatedByDescendants, replaceWith, startingFromIdInLineage) => reportsCreatedByDescendants.reduce((agg, doc) => {
-  if (lineageManipulation.replaceLineage(doc, 'contact', replaceWith, startingFromIdInLineage)) {
-    agg.push(doc);
-  }
-  return agg;
-}, []);
-
-const replaceLineageInContacts = (descendantsAndSelf, replacementLineage, contactId) => descendantsAndSelf.reduce((agg, doc) => {
-  const startingFromIdInLineage = doc._id === contactId ? undefined : contactId;
-  const parentWasUpdated = lineageManipulation.replaceLineage(doc, 'parent', replacementLineage, startingFromIdInLineage);
-  const contactWasUpdated = lineageManipulation.replaceLineage(doc, 'contact', replacementLineage, contactId);
-  if (parentWasUpdated || contactWasUpdated) {
-    agg.push(doc);
-  }
-  return agg;
-}, []);
-
-const replaceLineageInAncestors = (descendantsAndSelf, ancestors) => ancestors.reduce((agg, ancestor) => {
-  let result = agg;
-  const primaryContact = descendantsAndSelf.find(descendant => ancestor.contact && descendant._id === ancestor.contact._id);
-  if (primaryContact) {
-    ancestor.contact = lineageManipulation.createLineageFromDoc(primaryContact);
-    result = [ancestor, ...result];
-  }
-
-  return result;
-}, []);