11 KiB
| title | sidebar_label | pagination_prev | pagination_next |
|---|---|---|---|
| Sheets in PST Mailboxes | PST Mailboxes | demos/net/server/index | demos/net/headless |
import current from '/version.js'; import Tabs from '@theme/Tabs'; import TabItem from '@theme/TabItem'; import CodeBlock from '@theme/CodeBlock';
<head> </head>PST (Personal Storage Table) is a common file format for storing messages. Electronic discovery commonly involves extracting data from attached spreadsheets in e-mail messages stored in PST archives.
pst-extractor1 is a NodeJS module designed for extracting objects from PST
files. It has been used to extract spreadsheets from the Enron Corpus2 and
other large mailboxes.
SheetJS is a JavaScript library for reading and writing data from spreadsheets.
This demo uses pst-extractor and SheetJS to read spreadsheets. We'll explore
how to load SheetJS in a NodeJS script or website, extract spreadsheets files,
and generate HTML and CSV views of the underlying data.
The "Live Demo" reads PST files. Individual spreadsheets within the file can be downloaded or previewed in the browser.
:::note
This demo was last tested on 2023 October 22 against pst-extractor 1.9.0
:::
Overview
The SheetJS NodeJS module can be
imported from scripts that use pst-extractor.
Parsing PST Files
The pst-extractor module exposes a PSTFile class. The constructor requires a
proper NodeJS buffer.
The following snippet reads and parses enron.pst from the local filesystem.
fs.readFileSync3 accepts a filename and returns a Buffer:
const fs = require("fs"), PSTExtractor = require("pst-extractor");
const file = fs.readFileSync("enron.pst");
const pst = new (PSTExtractor.PSTFile)(file);
Walking the Tree
pst-extractor presents a tree-like structure to inspect the contents of the
PST file. It is recommended to use recursive functions to walk the tree.
The following tree walker will collect all XLSX and XLS attachments:
/* walk the PST file and add all attachments to the specified array */
function walk(f,arr) {
if(f.hasSubfolders) for(let sf of f.getSubFolders()) walk(sf,arr);
if(f.contentCount <= 0) return;
for(let e = f.getNextChild(); e != null; e = f.getNextChild()) {
for(let i = 0; i < e.numberOfAttachments; ++i) {
var a = e.getAttachment(i);
/* XLS spreadsheet test by filename */
if(/.xls[xmb]?$/.test(a.filename)) arr.push(a);
}
}
}
/* generate a list of attachments */
const files = [];
walk(pst.getRootFolder(), files);
Generating Buffers
The PSTAttachment class holds attachment metadata. To avoid loading everything
in memory, the raw data is exposed as a custom stream object. Since the SheetJS
read function requires data in a Buffer or Uint8Array, a helper function
is used to collect the data:
/* collect data from the attachment into a "Buffer" */
function collect(file) {
const strm = file.fileInputStream;
const data = Buffer.alloc(strm._length.low);
strm.readCompletely(data);
return data;
}
/* collect data from the first attachment */
const buf0 = collect(files[0]);
Processing Attachments
Given a NodeJS Buffer, the SheetJS read method4 parses the data and returns
a workbook object5. Individual worksheets can be extracted from the workbook
and converted to CSV6 or HTML7.
The following example prints the contents of each worksheet in CSV form:
const XLSX = require("xlsx");
/* parse workbook and print CSV contents of each sheet */
const wb = XLSX.read(buf0);
wb.SheetNames.forEach(n => {
const ws = wb.Sheets[n];
const csv = XLSX.utils.sheet_to_csv(ws);
console.log(`#### ${file.filename} ! ${n}`);
console.log(csv);
});
Browser Caveats
The SheetJS Standalone scripts
can be loaded through a SCRIPT tag.
This demo uses a special pst-extractor build for the web.
Compared to the NodeJS build, browser scripts require special Buffer wrappers.
For example, the following function will fail since the library does not support
ArrayBuffer objects:
async function error_fetch_and_parse_pst(url) {
const ab = await (await fetch(url)).arrayBuffer();
// this will throw an error
return new (PSTExtractor.PSTFile)(ab);
}
The browser build exposes the Buffer object in the PSTExtractor global:
async function correct_fetch_and_parse_pst(url) {
const ab = await (await fetch(url)).arrayBuffer();
// highlight-next-line
const buf = new PSTExtractor.Buffer(ab);
return new (PSTExtractor.PSTFile)(buf);
}
Browser Build
The pst-extractor library is designed for NodeJS. Parts of the library expect
a NodeJS Buffer, which does not exist in the browser. A fake Buffer can be
added and exposed in a script.
pstextractor.js is loaded in the demo page.
Build instructions (click to show)
- Initialize a new NodeJS project and install the dependency:
mkdir pstextract
cd pstextract
npm init -y
npm i --save pst-extractor@1.9.0
- Save the following to
shim.js:
const PSTExtractor = require("pst-extractor");
module.exports = PSTExtractor;
module.exports.Buffer = Buffer;
- Build the script:
npx browserify@17.0.0 -s PSTExtractor -o pstextractor.js shim.js
Demos
NodeJS
This demo will fetch a test PST and extract all embedded spreadsheets. The script can be adapted to read local PST files or pull PST files from a different URL.
- Initialize a new project:
mkdir sheetjs-pst
cd sheetjs-pst
npm init -y
- Install the SheetJS NodeJS module and
pst-extractor:
- Download
SheetJSPST.jsinto project folder:
curl -LO https://docs.sheetjs.com/pst/SheetJSPST.js
- Run the script:
node SheetJSPST.js
The process will fetch the test PST and extract the embedded spreadsheets. The terminal will display info on the exported files.
:::note pass
Lines starting with saving file show how attachments correspond to files. The
following line states that the first attachment (index 0) was originally named
RedRockA.xls and was saved to file0.xls on the file system:
saving file 0 |RedRockA.xls| to file0.xls
Lines starting with #### show the attachment file name and the worksheet name.
The following line explains that there is a worksheet named "Oct 26, 2001" in
the file RedRockA.xls:
#### RedRockA.xls ! Oct 26, 2001
Every other line is a CSV row from the named worksheet. For example, the first
four lines of worksheet "Oct 26, 2001" in RedRockA.xls are shown below:
#### RedRockA.xls ! Oct 26, 2001
// highlight-start
RED ROCK EXPANSION PROJECT,,,,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,,,
,,,, , , ,,,,,,,,,,,,
SHIPPER,CONTRACT #,Term,MMBtu/d,RECEIPT POINT,DELIVERY POINT,MMBtu/d,,,,,,,,,,,,
// highlight-end
:::
Live Demo
This demo reads PST mailboxes. Due to browser limitations, PST files larger than 100 MB may crash the browser.
After parsing the PST file, the "Attachments" table will list attached XLSX and XLS spreadsheets in the file. The "preview" link will display a HTML table with the data in the spreadsheet. The "download" link will download the attachment.
The test file was based on the EDRM clean extract from the "Enron Corpus" and includes a few XLS attachments.
:::caution pass
If the live demo shows a message
Please reload the page
please refresh the page. This is a known bug in the documentation generator.
:::
function SheetJSPreviewPSTSheets() {
const [ files, setFiles ] = React.useState([]);
const [ __html, setHTML ] = React.useState("");
/* recursively walk PST and collect attachments */
const walk = (f,arr) => {
if(f.hasSubfolders) for(let sf of f.getSubFolders()) walk(sf,arr);
if(f.contentCount <= 0) return;
for(let e = f.getNextChild(); e != null; e = f.getNextChild()) {
for(let i = 0; i < e.numberOfAttachments; ++i) {
var a = e.getAttachment(i);
/* XLS spreadsheet test by filename */
if(/.xls[xmb]?$/.test(a.filename)) arr.push(a);
}
}
}
/* collect data from the attachment into a "Buffer" */
const collect = (j) => {
const strm = files[j].fileInputStream;
const data = new PSTExtractor.Buffer(strm._length.low);
strm.readCompletely(data);
return data;
}
/* view selected attachment */
const view = (j) => {
const data = collect(j);
/* parse */
const wb = XLSX.read(data);
/* convert first sheet to HTML */
const ws = wb.Sheets[wb.SheetNames[0]];
setHTML(XLSX.utils.sheet_to_html(ws));
}
/* process array buffer */
const process_ab = (ab) => {
const pst = new (PSTExtractor.PSTFile)(new PSTExtractor.Buffer(ab));
const data = [];
walk(pst.getRootFolder(), data);
setFiles(data);
};
/* on click, fetch and process file */
const doit = async() => {
const ab = await (await fetch("/pst/enron.pst")).arrayBuffer();
process_ab(ab);
};
const chg = async(e) => process_ab(await e.target.files[0].arrayBuffer());
/* download selected attachment */
const dl = (j) => {
const a = document.createElement("a");
a.download = files[j].filename;
a.href = URL.createObjectURL(new Blob([collect(j)]));
document.body.appendChild(a);
a.click();
document.body.removeChild(a);
}
if(typeof PSTExtractor == "undefined") return <b>Please reload the page</b>;
return ( <>
<p>Use the file input to select a file, or click "Use a Sample PST"</p>
<input type="file" accept=".pst" onChange={chg}/>
<button onClick={doit}>Use a Sample PST!</button><br/><br/>
<table><thead><th colspan="3">Attachments</th></thead>
<tbody>{files.map((f,j) => (
<tr key={j}><th>{f.filename}</th>
<td><a onClick={()=>view(j)}>(preview)</a></td>
<td><a onClick={()=>dl(j)}>(download)</a></td>
</tr>
))}</tbody>
</table>
<b>Preview of first worksheet</b><br/>
<div dangerouslySetInnerHTML={{__html}}></div>
</> );
}
-
The project has no official website. The official repository is hosted on GitHub. ↩︎
-
Extracted spreadsheets are available on GitHub ↩︎
-
See
fs.readFileSyncin the NodeJS documentation ↩︎ -
See "Workbook Object" ↩︎