Started the reorganization of Scripts for github. Some notes changes.

Scripts/Auto_Scrape_Screenshots/combine_to_pdf.py

@@ -0,0 +1,71 @@
+from PIL import Image
+import glob
+import re
+import os
+from datetime import date
+
+currentdir = "/Users/normrasmussen/Documents/Work/Scripts/Walmart_Screenshots/"
+
+
+def find_pictures(currentdir):
+    files = []
+    listfiles = glob.glob(currentdir + "*.png")
+    for file in listfiles:
+        files.append(os.path.basename(file))
+        # Now file will only show the file name, not the entire path
+    split_resources(files, currentdir)
+
+
+def split_resources(files, currentdir):
+    try:
+        resource_title = files[0]
+        resource_title = resource_title[:-6]
+        files.sort()
+        new_list = []
+        for file in files:
+            if resource_title in file:
+                new_list.append(file)
+        for item in new_list:
+            files.remove(item)
+        split_resources(files, currentdir)
+        process_pictures(new_list, resource_title, currentdir)
+    except IndexError as e:
+        print(e)
+    finally:
+        pass
+
+
+def process_pictures(new_list, resource_title, currentdir):
+    resource_title = re.sub(r'[?]', "", resource_title)
+    today = date.today()
+    today = today.strftime("%m.%d.%Y")
+    image_list = []
+    resource = Image.open(new_list[0])
+    resource = resource.convert("RGB")
+    for picture in new_list[1:]:
+        image = Image.open(picture)
+        converted = image.convert("RGB")
+        image_list.append(converted)
+        # image_list.append(image)
+    resource.save(
+        currentdir + f"PDFs/{resource_title}_{today}.pdf",
+        save_all=True,
+        append_images=image_list,
+    )
+
+
+def delete_originals(currentdir):
+    path = glob.glob(currentdir + "*.png")
+    for file in path:
+        try:
+            os.remove(file)
+        except TypeError as e:
+            print("Error!")
+            print(e)
+        finally:
+            print("All Done")
+
+
+if __name__ == "__main__":
+    find_pictures(currentdir)
+    delete_originals(currentdir)

Scripts/Auto_Scrape_Screenshots/resources.js

@@ -0,0 +1,94 @@
+const puppeteer = require('puppeteer');
+const path = require('path')
+
+/*
+for (let i=0; i<urls.length; i++){
+  const url = urls[i];
+  await page.goto('${url}');
+  await page.waitForNavigation({ waitUntil: 'networkkidle2' });
+}
+
+const url = [
+  "https://walmart.northpass.com/app/courses/90bbdf4d-c674-403c-93d1-1c2e43364e0e", // Tax Filling FAQs
+  "https://walmart.northpass.com/app/courses/6b74d176-6027-40f5-af0c-fdee6d4f53bf", // Tips for Claiming your Branch Wallet
+  "https://walmart.northpass.com/app/courses/5624342f-9927-4704-bc6b-5c84068c0564", // Branch Wallet FAQs
+  "https://walmart.northpass.com/app/courses/65a83b52-08e9-48ac-b953-59dfa22f7358", // Pointers on Viewing your Earnings
+  "https://walmart.northpass.com/app/courses/0cd301a1-d98e-4837-ae9a-6fbd6a152aa3", // Curious how trip earnings works?
+  "https://walmart.northpass.com/app/courses/71a2064f-600d-415b-9b93-0a695793a09b", // Having Issues Confirming your arrival?
+  "https://walmart.northpass.com/app/courses/f9fea960-9fb3-4d05-858a-ea9a5657e20a", // Not Receiving App Notifications?
+  "https://walmart.northpass.com/app/courses/dce79912-ffe2-4648-a258-d725dbea6dc4", // Pointers on vehicle details
+  "https://walmart.northpass.com/app/courses/59a21c93-36ed-4e25-a723-6c8b3e6a80f0", // Pointers on Creating Your Account
+  "https://walmart.northpass.com/app/courses/17224d43-5acc-4dfb-805b-53f85cafb633", // Pointers on Downloading the App
+  "https://walmart.northpass.com/app/courses/bb1449eb-9b8d-48b2-9e05-1261171bd0e7", // Reviewing your trips
+  "https://walmart.northpass.com/app/courses/b30329fa-825f-4447-85f2-6caeea998ec6", // Pointers on Sharing Your Location
+  "https://walmart.northpass.com/app/courses/6c94e20c-aff7-465a-baaf-c32804fa7c9c", // Turning off Spark Now
+  "https://walmart.northpass.com/app/courses/2e616785-7ec7-40f0-a6a7-069f3c8c4c02", // Updating Spark Driver App
+  "https://walmart.northpass.com/app/courses/f3be53c0-cd21-4a2d-97f0-8b84ba030f4a", // Questions about Earnings
+  "https://walmart.northpass.com/app/courses/e889b1de-6305-424c-b0cd-47e36e798fa3", // Damaged Order or Item
+  "https://walmart.northpass.com/app/courses/8e20a3e6-37ef-4538-9047-4bcce5666d59", // Error messages & Alerts
+]
+*/
+const apiKey = "6hUfJdAartHTHhHc0WIRZYPWe"
+
+const url = process.argv[2];
+if (!url) {
+  throw "Please provide a URL as the first argument";
+}
+
+const uid = "/\?uid\=7b84eae4-fb34-4689-9f33-24071c3e5a41";
+const activity = new Array();
+const activityLinks = new Array();
+var num = 1
+
+/*
+function loopUrls() {
+  for (item of url){
+    console.log(item)
+    courseOverview()
+  }
+}*/
+
+async function courseOverview() {
+  const browser = await puppeteer.launch();
+  const page = await browser.newPage();
+  const course = url+uid;
+  await page.setViewport({ width:390, height:844 })
+  await page.goto(course);
+
+  await page.title();
+  const [getXpath] = await page.$x('/html/body/div[2]/div/div');
+  // const [getXpath] = await page.$x('//*[@id="course-desktop"]/div/div[3]/div[2]/div[2]/div[2]/ol/li/div');
+  const resourcetitle = await page.evaluate(name => name.innerText, getXpath);
+  const resource = resourcetitle.trim();
+  console.log(resource);
+
+  await page.screenshot({path: `${resource}_1.png`, fullPage:true, headless:false});
+
+  const hrefs = await page.$$eval('a', as => as.map(a => a.href));
+  var links = Object.entries(hrefs);
+  links.forEach(([key, value]) => {
+    var link = value;
+    if (link.includes("activities")) {
+      activity.push(link);
+    }
+  });
+  uniqueLinks = [...new Set(activity)];
+  for (const link of uniqueLinks) {
+    num = num+1
+    const newlink = link+uid;
+    // console.log(newlink)
+    screenshots(newlink, resource, num);
+  }
+  await browser.close();
+}
+
+async function screenshots(newlink, resource, num) {
+  const browser = await puppeteer.launch();
+  const page = await browser.newPage();
+  await page.setViewport({ width: 390, height: 844 })
+  await page.goto(newlink);
+  await page.screenshot({path: `${resource}_${num}.png`, fullPage: true, headless: false});
+  await browser.close();
+}
+
+courseOverview();

Scripts/Auto_Scrape_Screenshots/resources_api.js

@@ -0,0 +1,122 @@
+const puppeteer = require('puppeteer');
+const axios = require('axios')
+const path = require('path')
+
+const apiKey = "6hUfJdAartHTHhHc0WIRZYPWe"
+// Spark Production ^
+// const apiKey = "p5fidpuedHaOlPnd8EcpxzQMG"
+// Luminate Production ^
+const uid = "/\?uid\=7beg87y4-fh24-4929-3rt5-24kdn87s5241";
+const groupIds = []
+
+const getAllGroups = async (num) => {
+    if(num === 1){
+    }
+
+    let page = num;
+
+    await axios({
+        method: 'get',
+	url: 'https://api2.northpass.com/v2/courses?page=${page}&limit=200',
+        headers: {
+            'accept': '*/*',
+            'x-api-key': apiKey,
+            'content-type': 'application/json'
+        }
+    })
+    .then(async (res) => {
+        if (res.data.links.next != null) {
+            page++;
+            for (let i = 0; i < res.data.data.length; i++) {
+	      if (res.data.data[i].attributes.status == "live") {
+		console.log(res.data.data[i].attributes.name)
+		groupIds.push(res.data.data[i].id)
+	      }
+	    }
+            //await getAllGroups(page);
+        } else {
+           // for (let i = 0; i < res.data.data.length; i++) {
+	   //   if (res.data.data[i].attributes.status == "live") {
+	//	console.log(res.data.data[i].attributes.name)
+	//	groupIds.push(res.data.data[i].id)
+	  //    }
+//	}
+      }
+    })
+    .catch(err => {
+        console.log(err);
+    })
+  let i = 0
+  loopIds(i)
+  console.log(groupIds)
+}
+
+function loopIds(i) {
+  console.log("Loop Ids Function")
+  var num = 1
+  id = groupIds[i]
+  courseOverview(id, i, num)
+}
+
+async function courseOverview(id, i, num) {
+  console.log("Course Overview Function")
+  const activity = new Array();
+  const url = "https://walmart.northpass.com/app/courses/";
+  const browser = await puppeteer.launch();
+  const page = await browser.newPage();
+  const course = url+id+uid;
+  console.log(course)
+  if (course.includes("undefined")) {
+    console.log("Error - Undefined UUID. Possibly end of list. Exiting.")
+    await browser.close();
+  } else {
+      await page.setViewport({ width:390, height:844 })
+      await page.goto(course, {
+	    waitUntil: 'load',
+	    timeout: 0 });
+
+      await page.title();
+      const [getXpath] = await page.$x('/html/body/div[1]/div');
+      const resourcetitle = await page.evaluate(name => name.innerText, getXpath);
+      console.log(resourcetitle)
+      const resource = resourcetitle.trim();
+      console.log(resource);
+
+      await page.screenshot({path: `${resource}_1.png`, fullPage:true, headless:false});
+
+      const hrefs = await page.$$eval('a', as => as.map(a => a.href));
+      var links = Object.entries(hrefs);
+      links.forEach(([key, value]) => {
+	var link = value;
+	if (link.includes("activities") && (link.includes(id))) {
+	  console.log(link);
+	  activity.push(link);
+	}
+      });
+      setTimeout(() => {
+
+      }, 5000);
+      uniqueLinks = [...new Set(activity)];
+      for (const link of uniqueLinks) {
+	num = num+1
+	const newlink = link+uid;
+	await page.setViewport({ width: 390, height: 844 })
+	await page.goto(newlink, {
+	  waitUntil: 'load',
+	  timeout: 0 });
+	await page.screenshot({path: `${resource}_${num}.png`, fullPage: true, headless: false});
+	//screenshots(newlink, resource, num, i);
+	}
+      await browser.close();
+      i++;
+      loopIds(i);
+  }
+}
+
+async function screenshots(newlink, resource, num, i) {
+  const browser = await puppeteer.launch();
+  const page = await browser.newPage();
+  await browser.close();
+}
+
+getAllGroups();

Scripts/Auto_Scrape_Screenshots/walmart_screenshots.sh

@@ -0,0 +1,15 @@
+#! /bin/zsh
+
+jscript=$(node resources_api.js)
+render_pdf=$(python3 combine_to_pdf.py)
+
+echo "Loading virtual environment for Python"
+. ${PYENV_ROOT}/versions/walmart/bin/activate | awk 'NR==1{print $1,$2,$3}'
+
+echo "First up, Javascript's Pupeteer."
+echo "$jscript"
+
+echo "JS Complete. Running Python Script to clean up PDFs."
+echo "$render_pdf"
+
+echo "PDFs complete. Shutting down Virtual env."