sec-edgar-api 0.5.12 → 1.0.0

package/build/services/DocumentParser/parsers/parse-insider-transaction.d.ts

@@ -1,4 +1,4 @@
-import type { InsiderTransaction, Issuer, Owner, XMLParams } from '../../../types';
+import { Owner, InsiderTransaction, Issuer, XMLParams } from '../../../types';
 export declare function parseInsiderTransactions(params: XMLParams): {
     transactions: InsiderTransaction[];
     owners: Owner[];

package/build/services/DocumentParser/parsers/parse-insider-transaction.js

@@ -1,50 +1,49 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseInsiderTransactions = void 0;
-var HtmlTableExtractor_1 = require("../../HtmlTableExtractor");
+function parseTable(html) {
+    var rows = html.split('<tr');
+    return rows
+        .map(function (rowHtml) {
+        return rowHtml
+            .split('<td')
+            .map(function (cellHtml) { return cellHtml.substring(cellHtml.indexOf('>') + 1, cellHtml.lastIndexOf('</td>')); })
+            .slice(1);
+    })
+        .filter(function (row) { return row.length > 0; });
+}
+function stripHtml(html) {
+    return html
+        .replace(/<.*?>/g, '')
+        .replace(/&nbsp;/g, ' ')
+        .replace(/\n/g, ' ')
+        .replace(/\s+/, ' ')
+        .trim();
+}
+function toNumber(value) {
+    return Number(value.replace(/,/g, '').trim()) || 0;
+}
 function parseInsiderTransactions(params) {
-    var _a, _b, _c, _d, _e, _f, _g, _h, _j, _k, _l, _m, _o, _p, _q;
+    var _a, _b, _c, _d;
     var xml = params.xml;
-    var parser = new HtmlTableExtractor_1.default();
-    var tables = parser.extractTables(xml, {
-        stripHtml: true,
-        tagsToExclude: ['sup'],
-        stripParenthesis: true,
-        removeEmptyColumns: false,
-        getHeaderRowIndex: function (data) {
-            return data.rows.findIndex(function (row) {
-                var isNotEmptyRow = row.some(function (cell) { return cell.html.replace(/<.*?>/g, '').replace(/&.*?;/g, '').replace(/\s/g, '').length > 0; });
-                return isNotEmptyRow;
-            });
-        },
-    });
-    var getTableHeadHtml = function (table) {
-        var _a;
-        var match = table.html.replace(/\n/g, '').match(/<b>.*?<\/b>/g);
-        var matches = Array.isArray(match) ? match : [];
-        return (_a = matches.find(function (match) { return match.includes('<a'); })) !== null && _a !== void 0 ? _a : null;
-    };
-    var tableCompany = tables.find(function (table) { return table.html.includes('cgi-bin/browse-edgar?action=getcompany') && getTableHeadHtml(table) !== null; });
-    var tableCompanyHead = tableCompany ? getTableHeadHtml(tableCompany) || '' : '';
-    var issuerCik = Number((_b = (_a = tableCompanyHead.split('</a>')[0]) === null || _a === void 0 ? void 0 : _a.split('>').pop()) === null || _b === void 0 ? void 0 : _b.trim()) || 0;
-    var issuerName = ((_d = (_c = tableCompanyHead.split('(')[0]) === null || _c === void 0 ? void 0 : _c.split('>').pop()) === null || _d === void 0 ? void 0 : _d.trim()) || '';
-    var cells = (_e = tableCompany === null || tableCompany === void 0 ? void 0 : tableCompany.rows.flat()) !== null && _e !== void 0 ? _e : [];
-    cells.find(function (cell) { return cell.html.toLowerCase().includes('<b'); });
-    var getHeaderRow = function (table) { return table.rows.find(function (row) { return row.some(function (cell) { return cell.isHeaderRowCell; }); }); };
-    var findTableWithCol = function (colTextLower) {
-        return tables.find(function (table) { var _a; return (_a = getHeaderRow(table)) === null || _a === void 0 ? void 0 : _a.some(function (cell) { return String(cell.valueParsed).toLowerCase().includes(colTextLower); }); });
-    };
-    var tableOwners = findTableWithCol('type of owner');
-    var tableTransactions = findTableWithCol('security name');
-    var stripHtml = function (html) {
-        return html
-            .replace(/<.*?>/g, '')
-            .replace(/&nbsp;/g, ' ')
-            .replace(/\n/g, ' ')
-            .replace(/\s+/, ' ')
-            .trim();
-    };
-    var ownerByCik = new Map();
+    var ownerTableHeadingIndex = xml.indexOf('sortid=');
+    var transactionTableIdIndex = xml.indexOf('id="transaction-report"');
+    if (ownerTableHeadingIndex === -1) {
+        throw new Error('Owner table heading not found in XML data');
+    }
+    var ownerTableStartIndex = xml.lastIndexOf('<table', ownerTableHeadingIndex);
+    var ownerTableEndIndex = xml.indexOf('</table>', ownerTableHeadingIndex) + 8;
+    var ownerTableHtml = xml.substring(ownerTableStartIndex, ownerTableEndIndex);
+    var transactionTableStartIndex = xml.lastIndexOf('<table', transactionTableIdIndex);
+    var transactionTableEndIndex = xml.indexOf('</table>', transactionTableIdIndex) + 8;
+    var transactionTableHtml = xml.substring(transactionTableStartIndex, transactionTableEndIndex);
+    var issuerUrlIndex = xml.indexOf('cgi-bin/browse-edgar?action=getcompany');
+    var issuerUrlStartIndex = xml.lastIndexOf('<b>', issuerUrlIndex) + 3;
+    var issuerUrlEndIndex = xml.indexOf('</b>', issuerUrlIndex);
+    var issuerHtml = xml.substring(issuerUrlStartIndex, issuerUrlEndIndex).trim();
+    var issuerHtmlCikStart = issuerHtml.lastIndexOf('(') + 1;
+    var issuerCik = toNumber(stripHtml(issuerHtml.substring(issuerHtmlCikStart, issuerHtml.lastIndexOf(')'))));
+    var issuerName = issuerHtml.substring(0, issuerHtmlCikStart - 1).trim();
     var headerOwner = ['owner_name', 'filings', 'transaction_date', 'type_of_owner'];
     var headerTransaction = [
         'acquisition_or_disposition',
@@ -60,33 +59,37 @@ function parseInsiderTransactions(params) {
         'owner_cik',
         'security_title',
     ];
-    for (var _i = 0, _r = (_f = tableOwners === null || tableOwners === void 0 ? void 0 : tableOwners.rows) !== null && _f !== void 0 ? _f : []; _i < _r.length; _i++) {
-        var row = _r[_i];
-        if ((_g = row[0]) === null || _g === void 0 ? void 0 : _g.isHeaderRowCell)
-            continue;
+    var ownerByCik = new Map();
+    var ownerRows = parseTable(ownerTableHtml);
+    var transactionRows = parseTable(transactionTableHtml);
+    var isSwitchedOwnerIssuer = stripHtml((_b = (_a = ownerRows[0]) === null || _a === void 0 ? void 0 : _a[0]) === null || _b === void 0 ? void 0 : _b.toLowerCase()).includes('issuer') || false;
+    for (var i = 1; i < ownerRows.length; i++) {
+        var row = ownerRows[i];
+        // if (row.length === 0 || row[0].includes('Owner Name') || row[0].includes('Issuer')) continue // Skip header row
         var owner = {
             ownerName: '',
             ownerCik: 0,
             ownerPosition: null,
             isDirector: false,
             isOfficer: false,
+            isTenPercentOwner: false,
         };
-        for (var _s = 0, row_1 = row; _s < row_1.length; _s++) {
-            var cell = row_1[_s];
-            var colName = headerOwner[cell.colIndex];
-            var htmlStripped = stripHtml(cell.html);
+        for (var i_1 = 0; i_1 < row.length; i_1++) {
+            var colName = headerOwner[i_1];
+            var htmlStripped = stripHtml(row[i_1]);
             switch (colName) {
                 case 'owner_name':
                     owner.ownerName = htmlStripped.split('Current Name')[0];
                     break;
                 case 'filings':
-                    owner.ownerCik = Number(cell.valueParsed || 0) || 0;
+                    owner.ownerCik = toNumber(htmlStripped);
                     break;
                 case 'type_of_owner': {
-                    var parts = htmlStripped.split(':');
-                    owner.isDirector = (_h = parts[0]) === null || _h === void 0 ? void 0 : _h.toLowerCase().includes('director');
-                    owner.isOfficer = (_j = parts[0]) === null || _j === void 0 ? void 0 : _j.toLowerCase().includes('officer');
-                    owner.ownerPosition = ((_k = parts[1]) === null || _k === void 0 ? void 0 : _k.trim()) || null;
+                    var _e = htmlStripped.split(':').map(function (s) { return s.toLowerCase().trim(); }), ownerType = _e[0], _f = _e[1], position = _f === void 0 ? null : _f;
+                    owner.isDirector = ownerType.includes('director');
+                    owner.isOfficer = ownerType.includes('officer');
+                    owner.isTenPercentOwner = ownerType.includes('10 percent') || ownerType.includes('10%');
+                    owner.ownerPosition = position;
                     break;
                 }
             }
@@ -94,10 +97,10 @@ function parseInsiderTransactions(params) {
         ownerByCik.set(owner.ownerCik, owner);
     }
     var transactions = [];
-    for (var _t = 0, _u = (_l = tableTransactions === null || tableTransactions === void 0 ? void 0 : tableTransactions.rows) !== null && _l !== void 0 ? _l : []; _t < _u.length; _t++) {
-        var row = _u[_t];
-        if ((_m = row[0]) === null || _m === void 0 ? void 0 : _m.isHeaderRowCell)
-            continue;
+    for (var _i = 0, transactionRows_1 = transactionRows; _i < transactionRows_1.length; _i++) {
+        var row = transactionRows_1[_i];
+        if (row.length === 0 || row[0].includes('Acquisition or Dis'))
+            continue; // Skip header row
         var transaction = {
             ownerName: '',
             ownerCik: 0,
@@ -106,6 +109,9 @@ function parseInsiderTransactions(params) {
             issuerName: issuerName,
             isDirector: false,
             isOfficer: false,
+            accessionNumber: '',
+            deemedExecutionDate: '',
+            form: '',
             /** true = buy, false = sell */
             isAcquisition: false,
             isDirectOwnership: false,
@@ -115,17 +121,15 @@ function parseInsiderTransactions(params) {
             transactionShares: 0,
             sharesOwnedFollowingTransaction: 0,
             lineNumber: 0,
-            deemedExecutionDate: '',
-            form: '',
-            accessionNumber: '',
+            isTenPercentOwner: false,
         };
-        for (var _v = 0, row_2 = row; _v < row_2.length; _v++) {
-            var cell = row_2[_v];
-            var colName = headerTransaction[cell.colIndex];
-            var htmlStripped = stripHtml(cell.html);
+        for (var i = 0; i < row.length; i++) {
+            var colName = headerTransaction[i];
+            var html = row[i];
+            var htmlStripped = stripHtml(html);
             switch (colName) {
                 case 'acquisition_or_disposition':
-                    transaction.isAcquisition = String(cell.valueParsed).toLowerCase() === 'a';
+                    transaction.isAcquisition = String(htmlStripped).toLowerCase() === 'a';
                     break;
                 case 'transaction_date':
                     transaction.transactionDate = htmlStripped;
@@ -136,7 +140,7 @@ function parseInsiderTransactions(params) {
                 case 'reporting_owner':
                     break;
                 case 'form': {
-                    var url = (_p = (_o = cell.html.match(/href="([^"]+)"/)) === null || _o === void 0 ? void 0 : _o[1]) !== null && _p !== void 0 ? _p : '';
+                    var url = (_d = (_c = html.match(/href="([^"]+)"/)) === null || _c === void 0 ? void 0 : _c[1]) !== null && _d !== void 0 ? _d : '';
                     transaction.accessionNumber = url.substring(url.lastIndexOf('/') + 1, url.lastIndexOf('-'));
                     transaction.form = htmlStripped;
                     break;
@@ -148,22 +152,23 @@ function parseInsiderTransactions(params) {
                     transaction.isDirectOwnership = htmlStripped.toLowerCase().includes('d');
                     break;
                 case 'num_securities_transacted':
-                    transaction.transactionShares = Number(cell.valueParsed) || 0;
+                    transaction.transactionShares = toNumber(htmlStripped) || 0;
                     break;
                 case 'num_securities_following':
-                    transaction.sharesOwnedFollowingTransaction = Number(cell.valueParsed) || 0;
+                    transaction.sharesOwnedFollowingTransaction = toNumber(htmlStripped);
                     break;
                 case 'line_number':
-                    transaction.lineNumber = Number(cell.valueParsed) || 0;
+                    transaction.lineNumber = toNumber(htmlStripped);
                     break;
                 case 'owner_cik': {
-                    var owner = ownerByCik.get(Number(cell.valueParsed) || 0);
+                    var owner = ownerByCik.get(toNumber(htmlStripped) || 0);
                     if (owner) {
                         transaction.ownerName = owner.ownerName;
                         transaction.ownerCik = owner.ownerCik;
                         transaction.ownerPosition = owner.ownerPosition;
                         transaction.isDirector = owner.isDirector;
                         transaction.isOfficer = owner.isOfficer;
+                        transaction.isTenPercentOwner = owner.isTenPercentOwner;
                     }
                     break;
                 }
@@ -174,9 +179,6 @@ function parseInsiderTransactions(params) {
         }
         transactions.push(transaction);
     }
-    var isSwitchedOwnerIssuer = (_q = tableTransactions === null || tableTransactions === void 0 ? void 0 : tableTransactions.rows) === null || _q === void 0 ? void 0 : _q.some(function (r) {
-        return r.some(function (c) { return String(c.valueParsed).toLowerCase() === 'issuer'; });
-    });
     var owners = Array.from(ownerByCik.values());
     var issuers = [{ issuerName: issuerName, issuerCik: issuerCik }];
     // if searching by person, owner and issuer data will be switched
@@ -194,6 +196,10 @@ function parseInsiderTransactions(params) {
             transaction.ownerName = issuerName;
         });
     }
-    return { transactions: transactions, owners: owners, issuers: issuers };
+    return {
+        transactions: transactions,
+        owners: Array.from(ownerByCik.values()),
+        issuers: issuers,
+    };
 }
 exports.parseInsiderTransactions = parseInsiderTransactions;

package/build/types/insider-transaction.type.d.ts

@@ -5,6 +5,7 @@ export interface Owner {
     ownerPosition: string | null;
     isDirector: boolean;
     isOfficer: boolean;
+    isTenPercentOwner: boolean;
 }
 export interface Issuer {
     issuerName: string;
@@ -14,6 +15,7 @@ export interface InsiderTransaction {
     ownerName: string;
     ownerCik: number;
     ownerPosition: string | null;
+    isTenPercentOwner: boolean;
     issuerCik: number;
     issuerName: string;
     isDirector: boolean;
@@ -25,6 +27,7 @@ export interface InsiderTransaction {
     transactionDate: string;
     /**
      * ### Transaction Codes
+     * - P: Purchase
      * - S: Sale
      * - V: Voluntary Reporting
      * - A: Grant

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "sec-edgar-api",
-    "version": "0.5.12",
+    "version": "1.0.0",
     "description": "Fetch and parse SEC earnings reports and other filings. Useful for financial analysis.",
     "main": "build/index.js",
     "author": "Andrew Evers (https://github.com/andyevers)",