feat(pacer): Refine multi-document page handling logic

CHANGES.md

@@ -5,7 +5,7 @@
 The following changes are not yet released, but are code complete:
 
 Features:
- - None yet
+ - Refines logic to accurately identify multi-document pages containing a single file ([349](https://github.com/freelawproject/recap/issues/349), [402](https://github.com/freelawproject/recap-chrome/pull/402)) 
 
 Changes:
  - None yet

src/appellate/appellate.js

@@ -605,8 +605,15 @@ AppellateDelegate.prototype.handleAcmsDownloadPage = async function () {
           console.info(
             'RECAP: Got results from API. Processing results to insert banner'
           );
+          // To accurately identify ACMS documents, we should prioritize the
+          // `ACMS document details ID` stored in browser sessionStorage over
+          // the docket entry ID. This is because ACMS often uses the same URL
+          // for different attachments, making it ambiguous for identification
+          // purposes.
+          let acms_doc_id =
+            downloadData.docketEntryDocuments[0].docketDocumentDetailsId;
           let result = recapLinks.results.filter(
-            (obj) => obj.pacer_doc_id == this.docId,
+            (obj) => obj.acms_document_guid == acms_doc_id,
             this
           )[0];
           if (!result) return;
@@ -839,12 +846,13 @@ AppellateDelegate.prototype.attachRecapLinksToEligibleDocs = async function () {
   }
 
   // filter the links for the documents available on the page
-  let { links, docsToCases } = APPELLATE.findDocLinksFromAnchors(
-    this.links,
-    this.tabId,
-    this.queryParameters,
-    this.docketNumber
-  );
+  let { links, docsToCases, docsToAttachmentNumbers } =
+    APPELLATE.findDocLinksFromAnchors(
+      this.links,
+      this.tabId,
+      this.queryParameters,
+      this.docketNumber
+    );
 
   this.pacer_case_id = this.pacer_case_id
     ? this.pacer_case_id
@@ -864,6 +872,7 @@ AppellateDelegate.prototype.attachRecapLinksToEligibleDocs = async function () {
       caseId: this.pacer_case_id,
       docketNumber: this.docketNumber,
       docsToCases: docsToCases,
+      docsToAttachmentNumbers: docsToAttachmentNumbers,
     },
   });
 
@@ -1019,14 +1028,7 @@ AppellateDelegate.prototype.handleAttachmentPage = async function () {
   history.replaceState({ uploaded: true }, '');
 };
 
-AppellateDelegate.prototype.handleCombinedPdfPageView = async function () {
-  let warning = combinedPdfWarning();
-  document.body.appendChild(warning);
-};
-
-// If this page offers a single document, intercept navigation to the document
-// view page.
-AppellateDelegate.prototype.handleSingleDocumentPageView = async function () {
+AppellateDelegate.prototype.overrideDefaultForm = async function () {
   if (PACER.hasFilingCookie(document.cookie)) {
     let button = createRecapButtonForFilers(
       'Accept Charges and RECAP Document'
@@ -1050,6 +1052,60 @@ AppellateDelegate.prototype.handleSingleDocumentPageView = async function () {
   } else {
     await overwriteFormSubmitMethod();
   }
+};
+
+AppellateDelegate.prototype.handleCombinedPdfPageView = async function () {
+  // Find all `center` divs, which typically wrap receipt tables in lower
+  // courts. However, in appellate courts, these divs can also wrap the entire
+  // page content. To ensure an accurate count, we filter out nodes with more
+  // than 3 child elements, as a full page container would likely have more
+  // content.
+  let transactionReceiptTables = Array.from(
+    document.querySelectorAll('center')
+  ).filter((row) => row.childElementCount <= 3);
+  if (transactionReceiptTables.length === 0) return;
+
+  // In appellate courts, we don't rely on an exclusion list like lower courts.
+  // Instead, we extract document IDs from the URL parameters (`dls` attribute).
+  // These IDs represent the files included on the current page.
+  const urlParams = new URLSearchParams(window.location.search);
+  let includeList = urlParams.get('dls') ? urlParams.get('dls').split(',') : [];
+
+  // Count the number of receipt tables (from `transactionReceiptTables`) and
+  // documents listed in the URL (`includeList`). If either count is greater
+  // than 1, it indicates multiple documents are present. In this case, display
+  // a warning message.
+  if (
+    transactionReceiptTables.length > 1 ||
+    (includeList && includeList.length > 1)
+  ) {
+    const warning = combinedPdfWarning();
+    document.body.appendChild(warning);
+    return;
+  }
+
+  await this.overrideDefaultForm();
+
+  // When we receive the message from the above submit method, submit the form
+  // via XHR so we can get the document before the browser does.
+  window.addEventListener(
+    'message',
+    this.onDocumentViewSubmit.bind(this),
+    false
+  );
+
+  this.docId = await checkSingleDocInCombinedPDFPage(
+    this.tabId,
+    this.court,
+    this.docId,
+    true
+  );
+};
+
+// If this page offers a single document, intercept navigation to the document
+// view page.
+AppellateDelegate.prototype.handleSingleDocumentPageView = async function () {
+  await this.overrideDefaultForm();
 
   this.pacer_case_id = await APPELLATE.getCaseId(
     this.tabId,
@@ -1112,38 +1168,65 @@ AppellateDelegate.prototype.onDocumentViewSubmit = async function (event) {
   let form = document.getElementById(event.data.id);
 
   let title = document.querySelectorAll('strong')[1].innerHTML;
-  let dataFromTitle = APPELLATE.parseReceiptPageTitle(title);
+  let pdfData = APPELLATE.parseReceiptPageTitle(title);
 
-  if (!dataFromTitle.att_number && this.queryParameters.get('recapAttNum')) {
-    dataFromTitle.att_number = this.queryParameters.get('recapAttNum');
+  // For multi-document pages, the title alone doesn't provide sufficient
+  // information. Therefore, we need to extract additional data from the
+  // receipt table to accurately identify the PDF.
+  // Attempt to parse the necessary data from the receipt table.
+  if (!pdfData) pdfData = APPELLATE.parseDataFromReceiptTable();
+
+  // If we still don't have enough data after parsing the receipt table,
+  // submit the form without retrieving the file.
+  if (!pdfData) {
+    form.submit();
+    return;
   }
 
-  if (dataFromTitle.doc_number.length > 9) {
+  if (!pdfData.att_number) {
+    if (this.queryParameters.get('recapAttNum')) {
+      pdfData.att_number = this.queryParameters.get('recapAttNum');
+    } else {
+      pdfData.att_number = await getAttachmentNumberFromPacerDocId(
+        this.tabId,
+        this.docId
+      );
+    }
+  }
+
+  if (pdfData.doc_number.length > 9) {
     // If the number is really big, it's probably a court that uses
     // pacer_doc_id instead of regular docket entry numbering.
-    dataFromTitle.doc_number = PACER.cleanPacerDocId(dataFromTitle.doc_number);
+    pdfData.doc_number = PACER.cleanPacerDocId(pdfData.doc_number);
   }
 
-  if (!dataFromTitle) {
-    form.submit();
-    return;
-  }
   $('body').css('cursor', 'wait');
-  let query_string = new URLSearchParams(new FormData(form)).toString();
-  const resp = await window.fetch(form.action, {
-    method: form.method,
-    headers: {
+  // The form method in multi-document pages is a GET request, while
+  // single-document pages use a POST request. By checking the method here, we
+  // can reuse this code to retrieve the PDF file and display it appropriately
+  // for both page types.
+  let queryString = new URLSearchParams(new FormData(form));
+  let url = new URL(form.action);;
+  let method = form.method.toUpperCase();
+  let options = {};
+  if (method == 'GET') {
+    // If the method is GET, append query parameters to the URL
+    queryString.forEach((value, key) => url.searchParams.append(key, value));
+  } else {
+    options['method'] = method;
+    options['headers'] = {
       'Content-Type': 'application/x-www-form-urlencoded',
     },
-    body: query_string
-  });
+    options['body'] = queryString.toString();
+  }
+  const resp = await window.fetch(url, options);
   let helperMethod = handleDocFormResponse.bind(this);
   helperMethod(
     resp.headers.get('Content-Type'),
     await resp.blob(),
     null,
     previousPageHtml,
-    dataFromTitle
+    pdfData
   );
 };
 

src/appellate/utils.js

@@ -235,20 +235,23 @@ let APPELLATE = {
   findDocLinksFromAnchors: function (nodeList, tabId, queryParameters, docketNumber) {
     let links = [];
     let docsToCases = {};
+    let docsToAttachmentNumbers = {};
     Array.from(nodeList).map((a) => {
       if (!PACER.isDoc1Url(a.href)) return;
 
-      let docNum = PACER.getDocNumberFromAnchor(a) || queryParameters.get('recapDocNum');
+      let docNum =
+        PACER.getDocNumberFromAnchor(a) || queryParameters.get('recapDocNum');
       let doDoc = PACER.parseDoDocPostURL(a.getAttribute('onclick'));
-      if (doDoc && doDoc.doc_id && doDoc.case_id) {
-        docsToCases[doDoc.doc_id] = doDoc.case_id;
+      let pacerCaseId =
+        (doDoc && doDoc.case_id) || queryParameters.get('caseId');
+      if (doDoc && doDoc.doc_id && pacerCaseId) {
+        docsToCases[doDoc.doc_id] = pacerCaseId;
       }
 
       a.removeAttribute('onclick');
       a.setAttribute('target', '_self');
 
       let url = new URL(a.href);
-      let pacerCaseId = (doDoc && doDoc.case_id) || queryParameters.get('caseId');
       url.searchParams.set('caseId', pacerCaseId);
 
       if (docNum) {
@@ -288,11 +291,14 @@ let APPELLATE = {
       clonedNode.dataset.pacerCaseId = pacerCaseId;
       clonedNode.dataset.pacerTabId = tabId;
       clonedNode.dataset.documentNumber = docNum ? docNum : docId;
-      if (attNumber) clonedNode.dataset.attachmentNumber = attNumber;
+      if (attNumber) {
+        clonedNode.dataset.attachmentNumber = attNumber;
+        docsToAttachmentNumbers[docId] = attNumber;
+      }
 
       links.push(docId);
     });
-    return { links, docsToCases };
+    return { links, docsToCases , docsToAttachmentNumbers};
   },
 
   // get the docId from the servlet parameter of the attachment page or the single doc page
@@ -370,6 +376,24 @@ let APPELLATE = {
     return r;
   },
 
+  // returns data from the table of the Receipt Page as an object
+  parseDataFromReceiptTable: () => {
+    // this method uses regex expressions to match that information from the
+    // receipt table and returns an object with the following attributes:
+    //  - docket_number
+    //  - doc_number
+    //  - att_number (if applicable).
+    let search_string = $('td:contains(Case)').text();
+    let regex =
+      /Case: (?<docket_number>[^']*), Document: (?<doc_number>\d+)(-(?<att_number>\d+))?/;
+    let matches = regex.exec(search_string);
+    // If no matches were found, return null.
+    if (!matches) return null;
+    // If an attachment number was not found, set it to 0.
+    if (!matches.groups.att_number) matches.groups.att_number = 0;
+    return matches.groups;
+  },
+
   // Returns an object with the court Id and docket number core extracted from a link to district court
   getDatafromDistrictLinkUrl: (url) => {
     // Converts links to district courts like: