Add support for inline images

3 years ago · 44b3cfe6de
parent 1371daae7f
commit 44b3cfe6de
3 changed files with 86 additions and 18 deletions
--- a/mail-to-pdf.mjs
+++ b/mail-to-pdf.mjs
@ -1,5 +1,6 @@
 #!/usr/bin/zx --quiet

+var elparser = require("elparser");
 import { resolve, basename } from "path";

 const pwd = (await $`pwd`).stdout.replace("\n", "");
@ -37,22 +38,46 @@ let html =
        </body>
      </html>`;

+if (!html.includes("<body")) {
+  html = /* HTML */ ` <head>
+      <meta charset="utf-8" />
+    </head>
+    <body>
+      ${html}
+    </body>`;
+}
+
+if (
+  html.includes(
+    `<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-2">`
+  )
+) {
+  html = html.replace(
+    `<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-2">`,
+    ""
+  );
+}
+
+if (!html.includes(`<meta charset="utf-8"`)) {
+  html = html.replace("</head>", `<meta charset="utf-8" /></head>`);
+}
+
 const attachments = (await $`mu extract ${path}`).stdout;

-function getSimpleHeader(body, header_name) {
-  const regexp = new RegExp(`:${header_name} (.*)\n`);
-  console.log("get simple header", header_name, body);
-  return body.match(regexp)[1];
+async function emacsEval(expr) {
+  return (await $`emacsclient --eval ${expr}`).stdout.slice(0, -1); // slice to remove trailing newline;
 }

-function formatAddress(sexp) {
-  sexp = sexp
-    .replace(/(^\(\(|\)\)$)/g, "")
-    .replace(/"? \. "/, " <")
-    .replace(/"$/, ">")
-    .replace(/^"/, "")
-    .replace(/^nil <(.*)>$/, "$1");
-  return sexp;
+async function getSimpleHeader(body_sexp, header_name) {
+  return await emacsEval(`(plist-get '${body_sexp} ':${header_name})`);
+}
+
+async function formatAddress(sexp) {
+  const [name, address] = await Promise.all([
+    emacsEval(`(car (car '${sexp}))`),
+    emacsEval(`(cdr (car '${sexp}))`),
+  ]);
+  return `${name === "nil" ? "" : name} <${address}>`;
 }

 function escape(str) {
@ -120,14 +145,45 @@ html = html.replace(
    </ul></body>`
 );

-const subject = getSimpleHeader(mail_body, "subject").slice(1, -1); // slice to remove "
-const from = formatAddress(getSimpleHeader(mail_body, "from"));
-const to = formatAddress(getSimpleHeader(mail_body, "to"));
-const date = await formatDate(getSimpleHeader(mail_body, "date"));
+const subject = (await getSimpleHeader(mail_body, "subject")).slice(1, -1); // slice to remove " and newlines
+const from = await formatAddress(await getSimpleHeader(mail_body, "from"));
+const to = await formatAddress(await getSimpleHeader(mail_body, "to"));
+const date = await formatDate(await getSimpleHeader(mail_body, "date"));
+const parts = elparser
+  .parse1(`'${await getSimpleHeader(mail_body, "parts")}`)
+  .qsexp.list.map((entry) =>
+    Object.fromEntries(
+      entry.list.map((value, index, array) => {
+        if (index % 2 == 0) {
+          const key = array[index].symbol.slice(1);
+          let value = array[index + 1];
+          for (const subkey of ["string", "symbol"]) {
+            if (value[subkey] !== undefined) {
+              value = value[subkey];
+            }
+          }
+          if (value.type && value.type?.type === "int") {
+            value = parseInt(value.val);
+          }
+          return [key, value];
+        } else return [];
+      })
+    )
+  );
+
+const inline_attachments = parts.filter((e) => e.cid !== undefined);
+
+for (const inline_attachment of inline_attachments) {
+  await $`mu extract --overwrite ${path} ${inline_attachment.name} --target-dir=/tmp --overwrite`;
+  html = html.replace(
+    `cid:${inline_attachment.cid}`,
+    `./${inline_attachment.name}`
+  );
+}

 html = html.replace(
  /<body>/i,
-  `<meta charset="utf-8"><body>` +
+  `<body>` +
    /* HTML */ `<h1>${subject}</h1>
      <table>
        <tr>
@ -159,7 +215,7 @@ const output_file = `${resolve(path, "../")}/${basename(path)}.pdf`.replace(
  " "
 );

-await $`wkhtmltopdf ${output} ${output_file}`;
+await $`wkhtmltopdf  --enable-local-file-access ${output} ${output_file}`; // --enable-local-file-access  is there so the inline images don't have to be included as base64. Probably not the safest approach and base64 would be better
 // console.log(pwd + "output.html");
 // console.log(mail_body);

--- a/package-lock.json
+++ b/package-lock.json
@ -9,9 +9,15 @@
      "version": "1.0.0",
      "license": "ISC",
      "dependencies": {
+        "elparser": "^0.0.9",
        "sexpr-plus": "^7.0.0"
      }
    },
+    "node_modules/elparser": {
+      "version": "0.0.9",
+      "resolved": "https://registry.npmjs.org/elparser/-/elparser-0.0.9.tgz",
+      "integrity": "sha1-+bQxsVZSCK8roHKKE8Ux4FWbrSI="
+    },
    "node_modules/sexpr-plus": {
      "version": "7.0.0",
      "resolved": "https://registry.npmjs.org/sexpr-plus/-/sexpr-plus-7.0.0.tgz",
@ -19,6 +25,11 @@
    }
  },
  "dependencies": {
+    "elparser": {
+      "version": "0.0.9",
+      "resolved": "https://registry.npmjs.org/elparser/-/elparser-0.0.9.tgz",
+      "integrity": "sha1-+bQxsVZSCK8roHKKE8Ux4FWbrSI="
+    },
    "sexpr-plus": {
      "version": "7.0.0",
      "resolved": "https://registry.npmjs.org/sexpr-plus/-/sexpr-plus-7.0.0.tgz",
--- a/package.json
+++ b/package.json
@ -13,6 +13,7 @@
  "author": "",
  "license": "ISC",
  "dependencies": {
+    "elparser": "^0.0.9",
    "sexpr-plus": "^7.0.0"
  }
 }