From eb908dd8f00f3a5f44c0e535aa4a8b37ed44373a Mon Sep 17 00:00:00 2001
From: Lightling <contact@lightling.xyz>
Date: Fri, 9 Feb 2024 20:31:21 -0500
Subject: [PATCH] concurrent retrieval with user-limited threadcount

---
 buildDb.js  | 74 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 config.json | 26 +++++++++++++++++++
 getUser.js  | 34 ++++++++++++++++++++++++
 3 files changed, 134 insertions(+)
 create mode 100644 buildDb.js
 create mode 100644 config.json
 create mode 100644 getUser.js

diff --git a/buildDb.js b/buildDb.js
new file mode 100644
index 0000000..46fcbc1
--- /dev/null
+++ b/buildDb.js
@@ -0,0 +1,74 @@
+import { readFile } from 'fs/promises';
+
+import { getArg } from './args.js';
+import { getUser } from './getUser.js';
+import { error, log } from './log.js';
+
+const ctx = 'buildDb.js';
+
+const getMany = (processes, threadMax, directory) => new Promise((resolve, reject) => {
+  let running = 0;
+  let index = 0;
+
+  const get = () => {
+    const onFinish = (currentIndex) => {
+      log(ctx, `Finished ${processes[currentIndex].user}/media`);
+      --running;
+      get();
+    }
+
+    while (running < threadMax && index < processes.length) {
+      ++running;
+      let currentIndex = index++;
+
+      let proc = getUser(processes[currentIndex].user, directory);
+      proc.stdout.on('data', data => {
+        processes[currentIndex].logs.push(data);
+      });
+      proc.stderr.on('data', _ => onFinish(currentIndex));
+      proc.on('close', _ => onFinish(currentIndex));
+      proc.on('error', _ => onFinish(currentIndex));
+    }
+
+    if (running === 0) {
+      resolve();
+    }
+  }
+  get();
+});
+
+const buildDb = async () => {
+  log(ctx, 'Grabbing db');
+  let directory = '', threadMax = 1, db;
+  try {
+    directory = getArg('path');
+  } catch (err) {
+    error(ctx, err);
+    return;
+  }
+  try {
+    threadMax = getArg('threads');
+    log(ctx, `Using ${threadMax} threads`);
+  } catch (err) {
+    log(ctx, 'Using 1 thread');
+  }
+  try {
+    let file = await readFile(`${directory}/db.json`, { encoding: 'utf8' });
+    db = JSON.parse(file);
+  } catch (err) {
+    error(ctx, err);
+    return;
+  }
+
+  let processes = db.map(entry => ({
+    ...entry,
+    logs: [],
+  }));
+
+  log(ctx, `Building db using <user>/media for ${processes.length} users`);
+  await getMany(processes, threadMax, directory);
+
+  log(ctx, 'Building db using /search');
+}
+
+buildDb();
diff --git a/config.json b/config.json
new file mode 100644
index 0000000..07259fa
--- /dev/null
+++ b/config.json
@@ -0,0 +1,26 @@
+{
+  "extractor": {
+    "twitter": {
+      "base-directory": ".",
+      "directory": [
+        "{author['name']}"
+      ],
+      "filename": {
+        "locals().get('bitrate', 0) > 0": "{author['name']}-{tweet_id}-{date:Olocal/%Y%m%d_%H%M%S}-vid{num}.{extension}",
+        "locals().get('bitrate') == 0": "{author['name']}-{tweet_id}-{date:Olocal/%Y%m%d_%H%M%S}-gif{num}.{extension}",
+        "": "{author['name']}-{tweet_id}-{date:Olocal/%Y%m%d_%H%M%S}-img{num}.{extension}"
+      },
+      "cookies": "$HOME/.config/gallery-dl/twitter-cookies.txt",
+      "timeline": {
+        "strategy": "media"
+      },
+      "postprocessors": [
+        {
+          "name": "metadata",
+          "event": "post",
+          "filename": "{author[name]}_{tweet_id}-{date:Olocal/%Y%m%d_%H%M%S}.json"
+        }
+      ]
+    }
+  }
+}
diff --git a/getUser.js b/getUser.js
new file mode 100644
index 0000000..74ce257
--- /dev/null
+++ b/getUser.js
@@ -0,0 +1,34 @@
+import { spawn } from 'child_process';
+
+import { getArg } from './args.js';
+import { error, log } from './log.js';
+
+const ctx = 'getUser.js';
+
+export const getUser = (user, path) => {
+  const url = `https://twitter.com/${user}/media`;
+  let args;
+  try {
+    args = getArg('args');
+  } catch (err) {
+    log(ctx, 'No args being provided to gallery-dl');
+  }
+
+  log(ctx, `python3 ~/.local/bin/gallery-dl -c ./config.json${!!args ? ' ' + args + ' ' : ' '}-d ${path} ${url}`);
+  const proc = spawn(`python3 ~/.local/bin/gallery-dl -c ./config.json${!!args ? ' ' + args + ' ' : ' '}-d ${path} ${url}`, { shell: true });
+
+  proc.stdout.on('data', data => {
+    //log(ctx, data);
+  });
+  proc.stderr.on('data', data => {
+    error(ctx, data);
+  });
+  proc.on('error', err => {
+    error(ctx, err);
+  });
+  proc.on('close', code => {
+    log(ctx, `child process exited with code ${code}`);
+  });
+
+  return proc;
+};