refac to prepare database injestion

README.md

@@ -4,7 +4,7 @@
 
 ```sh
 pnpm install
-pnpm dev
+pnpm dev # for vue dev UI
 ```
 
 ## Import Cesium+ data
@@ -32,6 +32,14 @@ time npx tsx src/scripts/cesium-plus-import.ts
 # bafyreigczogsiuhaqus7eucalkwsy4vfkh3f4zg3c3rkvltxrwji6p5rnq
 ```
 
+## Start collector and indexer
+
+To start pubsub collector to IPFS and database indexer
+
+```sh
+npx tsx src/scripts/start-indexer.ts
+```
+
 ## TODO
 
 When using Kubo node, libp2p is not needed at all.

src/indexer/handlers.ts

+import type { DiffData } from './types'
+import { CID } from 'multiformats'
+
+// diff indexer event handler
+export async function indexDiff(diff: DiffData) {}
+
+// start indexer event handler
+export async function indexStart(cid: CID) {}

src/indexer/types.ts

+import { CID } from 'multiformats'
+import type { IndexRequest } from '../types'
+
+// data added between two CIDs
+export interface DiffData {
+  oldCID: CID
+  newCID: CID
+  newItems: Array<[CID, IndexRequest]>
+}

src/scripts/start-indexer.ts

@@ -6,17 +6,17 @@ import type { IndexRequest } from '../types'
 import { IPNS, EMPTY_NODE_CID } from '../consts'
 import { CID } from 'multiformats'
 import EventEmitter from 'events'
+import { indexDiff, indexStart } from '../indexer/handlers'
 
 // this script allows to start an indexer
 
+// === GLOBALS ===
+
 // queue of index requests waiting to be processed
 const processQueue: Array<[CID, IndexRequest]> = []
-const queueEvents = new EventEmitter()
+const events = new EventEmitter()
 let isProcessingQueue = false
 
-// show args
-// console.log(process.argv)
-
 /// global rootCID variable
 // initialize it:
 // - from CID if given
@@ -25,8 +25,23 @@ let isProcessingQueue = false
 // - as empty node else
 let rootCID = EMPTY_NODE_CID
 
+// === EVENTS ===
+
+// event type
+enum evtype {
+  // event to trigger collecting new index requests in queue
+  trigger = 'trigger',
+  // event to trigger injestion of new requests in database
+  indexDiff = 'indexThat',
+  // event to trigger database sync from scratch or last state
+  indexStart = 'start'
+}
+
+// === INIT ===
+
 // get root cid from arg if given
-if (process.argv.length >= 3) {
+async function setRootCIDfromArgs(argv: string[]) {
+  if (argv.length >= 3) {
     const arg = process.argv[2]
     try {
       // try parse as CID
@@ -59,10 +74,11 @@ if (process.argv.length >= 3) {
       console.log('using empty node insead')
     }
   }
+}
 
-// === define what to do of pubsub messages and start listening ===
+// === HANDLERS ===
 
-// message handler
+// pubsub message handler
 function handleMessage(message: any) {
   // console.log('received message')
   // console.log(message)
@@ -90,7 +106,7 @@ function handleMessage(message: any) {
         // low trust => sandboxed HAMT
 
         processQueue.push([cid, dag])
-        queueEvents.emit('trigger')
+        events.emit(evtype.trigger)
 
         // this is all that this indexer does
         // the rest (autopinning, postgres indexing... will be managed by an other part of the indexer)
@@ -108,11 +124,11 @@ function handleBatch() {
   // ignore event if already processing something or if queue is empty
   if (isProcessingQueue) return
   if (processQueue.length == 0) return
-  // if not processing do process
+  // if not processing, do process
   isProcessingQueue = true
   // take elements from queue
   let i = undefined
-  const items = []
+  const items: Array<[CID, IndexRequest]> = []
   while ((i = processQueue.shift()) != undefined) {
     items.push(i)
   }
@@ -126,22 +142,39 @@ function handleBatch() {
     .then((v) => mergeInodesSync(v.value, tree))
     .then((cid) => {
       // update CID and schedule publishing of new CID in history
+      const oldCID = rootCID
       rootCID = cid
       console.log(`new root CID ${cid}`)
       kubo.name.publish(rootCID, { ttl: '1s' })
       publishHistory(rootCID)
       isProcessingQueue = false
-      queueEvents.emit('trigger')
+      // trigger an other event in case new requests arrived meanwhile
+      events.emit(evtype.trigger)
+      // emit event to be processed by indexer
+      events.emit(evtype.indexDiff, { oldCID: oldCID, newCID: rootCID, newItems: items })
     })
 }
 
-queueEvents.on('trigger', handleBatch)
-queueEvents.on('trigger', handleBatch)
+// ===================== START ===========================
 
-// console.log(await kubo.pubsub.ls())
-// console.log(await kubo.pubsub.peers(TOPIC))
+// set root CID from CLI args
+await setRootCIDfromArgs(process.argv)
+
+// bind event handlers
+events.on(evtype.trigger, handleBatch)
+events.on(evtype.indexDiff, indexDiff)
+events.on(evtype.indexStart, indexStart)
 kubo.pubsub.subscribe(TOPIC, handleMessage)
 
+// emit event to tell indexer to start indexing to database
+// if it is starting from scratch, it will iterate over all values
+// if it already indexed up to a given cid, it will only iterate over the diff
+events.emit(evtype.indexStart, rootCID)
+// // optionally publish new cid and history with
+// kubo.name.publish(rootCID, { ttl: '1s' })
+// publishHistory(rootCID)
+
+// process loop
 console.log('listening...')
 setInterval(() => {}, 1 << 30)
 ;['SIGINT', 'SIGTERM', 'SIGQUIT'].forEach((signal) => process.on(signal, process.exit))