JavaScript Extractor Scripts

Kitinerary provides a JavaScript-engine for writing extractors to parse and output structured data from tickets and other travel documents. This data is later used in many projects to generate useful information for the user.

All extractor scripts are written in JavaScript and stored in /src/lib/scripts.

How to make your own extractor script

It's highly recommended to use the KItinerary Workbench to develop and test your extractor scripts. How to install and more information, see the KItinerary Workbench README.

Creating a new extractor script

To create a new extractor script, you need to create a new files in the $XDG_DATA_DIRS/kitinerary/extractors (~/ .local/share/kitinerary/extractors`) directory.

Note: For easier management and later collaboration in Git we recommend linking extractor scripts from the directory to a Git repository (ln -s $(pwd)/src/lib/scripts ~/.local/share/kitinerary/extractors).

Script declaration

Kitinerary uses a JSON file to declare the extractor scripts. This file sets filtering rules by which it knows, which extractor to run and defines the script itself.

Note: Multiple extractors can run on a single document, if more than one extractor outputs valid data it will be merged into single output.

Note: Multiple script declarations can exist for one js extractor file. Usefull if there many types of documents but same script can be used to extract data from them.

Extractor declaration

It contains the MIME type of the document that is going to be ingested, filter defining when this extractor script should run, and declaration script and the function that will be called for it.

{
  "mimeType": "application/pdf",
  // MIME type of the document that is going to be ingested (email or pdf, or other)
  "filter": [
    {
      ...
    }
  ],
  // Rules by which detect this ticket should be parsed by this script
  "script": "my-extractor-script.js",
  // Name of the extractor script itself
  "function": "extractTicket"
  // Name of the function in the script that will be called
}

Extractor scripts are run against a document node if all of the following conditions are met:

• The mimeType of the script matches that of the node.
• At least one of the extractor filter of the script match the node.

Extractor filters

Extractor filters are evaluated against document nodes (content of the document).

Extractor script filter consists of the following four properties:

{
  "[...]": "[...]",
  "filter": [
    {
      "mimeType": "text/plain",
      // Specifies the type of document you're looking for (e.g., QR code in document, plain text).
      "field": "",
      // If it is a complex document (like email with fields), it sets which filed of the document run "match" on. This is ignored for nodes containing basic types such as plain text or binary data.
      "match": "KDE.org/airlines",
      // A regular expression or exact string we look for.
      "scope": "Current"
      // Defines the relation to the node the script should be run on (Current, Parent, Children, Ancestors or Descendants).
    }
    // [...]
  ]
}

Scope defines where to match the filter in relation to mimeType. The following values are supported:

• Current: The filter is applied to the node itself (mimeType in filter).
• Parent: The filter is applied to the direct parent node of the current node (only one back).
• Children: The filter is applied to the direct child nodes of the current node (only one forward).
• Ancestors: The filter is applied to all parent nodes of the current node (all the way back).
• Descendants: The filter is applied to all child nodes of the current node (all the way forward).

Scope examples

---

We have an email with PDF, and ticket details inside the PDF. The PDF is a child of the email, and the ticket details are inside the PDF.

└── message/rfc822 // Email send by booking operator
    ├── text/html // HTML content of the email
    │   └── text/plain // extracted only text from the HTML
    ├── application/pdf // PDF attached to the email
    │   ├── internal/qimage // Image of the QR code
    │   │   └── text/plain // Decoded text from the QR code
    │   └── text/plain // Usually text inside the PDF
    └── internal/qimage // Image of booking company logo - we dont care 

{
  "mimeType": "application/pdf",
  // We are looking for PDF
  "field": "From",
  // Email (message/rfc822) has "From" field, with sender
  "match": "^booking@exampl-operator\.com$",
  "scope": "Parent"
  // We look at the parent of the PDF, which is the email
}

function parser(pdfTicket, node, matched) {
    [...]
}

Node[0] is the PDF, but we match ticket based on it's parrent which is Email, and we see if it was send from " booking@exampl-operator.com". Which results in the first argument of the parser function becoming the PDF, second argument is the node (PDF) and third argument is the matched the document (message/rfc822).

Examples

---

Anything attached to an email sent by "booking@example-operator.com". The field matched against here is the From header of the MIME message.

{
  "mimeType": "message/rfc822",
  // Its mime type of email
  "field": "From",
  // We look at fiels "From" in the email, which is the sender
  "match": "^booking@exampl-operator\.com$",
  // We look at exactly "booking@exampl-operator.com"
  "scope": "Ancestors"
}

---

Documents containing a barcode of the format "F12345678". Note that the scope here is Descendants rather than Children as the direct child nodes tend to be the images containing the barcode.

{
  "mimeType": "text/plain",
  // We look at plain text
  "scope": "Ancestors",
  "match": "^F\d{8}$"
  // We look for exactly "F" followed by 8 digits
}

---

Apple Wallet passes issued by "org.kde.travelAgency".

{
  "mimeType": "application/vnd.apple.pkpass",
  // We look at Apple Wallet passes
  "field": "passTypeIdentifier",
  // We look at field "passTypeIdentifier" which is the issuer
  "match": "org.kde.travelAgency",
  "scope": "Current"
  // We look only at this document
}

---

iCal events with an organizer email address of the "kde.org" domain. Note that the field here accesses a property of a property. This works at arbitrary depth, as long as the corresponding types are introspectable by Qt.

{
  "mimeType": "internal/event",
  "field": "organizer.email",
  "match": "@kde.org$",
  "scope": "Current"
}

---

A (PDF) document containing an IATA boarding pass barcode of the airline "AB". Triggering vendor-specific UIC or ERA railway tickets can be done very similarly, matching on the corresponding carrier ids.

{
  "mimeType": "internal/iata-bcbp",
  "field": "operatingCarrierDesignator",
  "match": "AB",
  "scope": "Descendants"
}

---

A node that has already existing results containing a reservation from "My Transport Operator". This is useful for scripts that want to augment or fix schema.org annotation already provided by the source. Note that the mimeType "application/ld+json" is special here as it doesn't only trigger on the document node content itself, but also matches against the result of nodes of any type.

{
  "mimeType": "application/ld+json",
  "field": "reservationFor.provider.name",
  "match": "My Transport Operator",
  "scope": "Current"
}

---

NOT RECOMMENDED This should be used as a last resort only, as matching against the full PDF document content can be expensive.

PDF documents containing the string "My Ferry Booking" anywhere.

{
  "mimeType": "application/pdf",
  "field": "text",
  "match": "My Ferry Booking",
  "scope": "Current"
}

Extractor script

Extractor scripts are run inside a QJSEngine, it isn't a full JS environment, and not everything is supported. There are some additional APIs available to extractor scripts (technical docs can be found here KItinerary::JsApi.

Objects of a document

ExtractorDocumentNode (node)

It's a object that represents a node in the document tree:

• content: Value of the node (eg. text, barcode content, etc)
• childNodes: List of child of this node, they are also ExtractorDocumentNode objects.
• mimeType: MIME type of the node (eg. text/plain, application/pdf, internal/qimage etc)

Examples

└── application/pdf // Ticket in PDF format
    ├── internal/qimage // Image of the QR code
    │   └── text/plain // Decoded text from the QR code
    └── text/plain // Usually text inside the PDF

function main(pdf, node) {
    cnsole.log(pdf.content); // Automagically extracted PDF content, no need to point at it.
    let imageOfQR = node.childNodes[0];
    let textFromQR = imageOfQR.childNodes[0].content;
}

DocumentNode types

Ticket itself can be in different formats, and each format has its own object:

PDF - PDF document

PdfDocument is a object that represents a PDF document; it has the following properties:

• text: Extracts text from the PDF page. If used on root node, it extracts all text from the PDF.
• pages: List of pages in the PDF
• textInRect: Extracts text from a given rectangle on the PDF page. Uses normalized coordinates (0-1) in format "Left, Top, Right, Bottom".

More: PdfDocument

Examples

// If ticket is in PDF the fist argument is the `PdfPage` object
function main(contentPDF, node) {
    const allText = contentPDF.text; // Extracts all text from the PDF page
    const firstPage = contentPDF.pages[0].text; // Extracts text from only from first page
    const textInRect = contentPDF.pages[0].textInRect(0, 0, 0.3, 0.25); // "Passanger: Kandalf"
}

Html - HTML document

HtmlDocument is an object that represents an HTML document consisting HtmlElements; it has the following properties and methods:

• rawData(): Returns the raw textual HTML data.
• root(): Returns the root element of the document.
• eval(xpath): Evaluates an XPath expression relative to the document root and returns matching elements.

HtmlElement represents an element within an HTML document; it has the following properties and methods:

• name: Returns the element name (tag).
• isNull: Checks if the element is null/invalid.
• attribute: Returns the value of the specified attribute.
• hasAttribute: Checks whether an attribute with the given name exists.
• attributes: Returns a list of all attributes of this element.
• content: Returns the immediate text content of this element (trimmed of whitespace).
• recursiveContent: Returns the text content of this element and all its children.
• parent: Returns the parent element of this node.
• firstChild: Returns the first child element of this node.
• nextSibling(: Returns the next sibling element of this node. -eval`: Evaluates an XPath expression relative to this element.

More: HtmlDocument, HtmlElement

Examples

  // Create a simple HTML document
const simpleHtml = `
    <html><head>
    <title>Flight Details</title>
    </head><body>
  
    <div class="flight-info">
    <h1>Flight KDE1996</h1>
    <div class="departure">
    <span class="code">KDQ</span>
    <span class="time">2025-02-20 08:30</span>
    </div>
  
    <div class="arrival">
    <span class="code">KDA</span>
    <span class="time">2025-02-22 16:45</span>
    </div>
  
    <div class="passenger" id="traveler">
    <span class="name">Kandalf the wizard</span>
    <span class="seat">12A</span>
  
    </div>
    </div>
    </body>
    </html>
    `;
 
const html = ExtractorEngine.extract(simpleHtml, "text/html").content;
 
const res = JsonLd.newFlightReservation();
 
// Get flight number from h1
const flightHeader = html.eval("//h1")[0];
console.log(flightHeader.content);
if (typeof flightHeader.content == "string") {
    const flightNumber = flightHeader.content.match(/Flight ([A-Z]{2})(\d+)/);
    if (flightNumber) {
        res.reservationFor.airline.iataCode = flightNumber[1];
        res.reservationFor.flightNumber = flightNumber[2];
    }
}
 
// Get departure info
const departureElement = html.eval("//div[@class='departure']")[0];
if (typeof flightHeader.content == "string") {
    const codeElement = departureElement.eval("span[@class='code']")[0];
    const timeElement = departureElement.eval("span[@class='time']")[0];
 
    res.reservationFor.departureAirport.iataCode = codeElement.content;
    res.reservationFor.departureTime = JsonLd.toDateTime(
        timeElement.content,
        "yyyy-MM-dd HH:mm",
        "en",
    );
}
 
// Get arrival info
const arrivalElement = html.eval("//div[@class='arrival']")[0];
if (typeof flightHeader.content == "string") {
    const codeElement = arrivalElement.eval("span[@class='code']")[0];
    const timeElement = arrivalElement.eval("span[@class='time']")[0];
 
    res.reservationFor.arrivalAirport.iataCode = codeElement.content;
    res.reservationFor.arrivalTime = JsonLd.toDateTime(
        timeElement.content,
        "yyyy-MM-dd HH:mm",
        "en",
    );
}
 
// Get passenger info using element navigation
const passengerDiv = html.eval("//div[@id='traveler']")[0];
const nameSpan = passengerDiv.firstChild;
const seatSpan = nameSpan.nextSibling;
 
res.underName = {
    "@type": "Person",
    name: nameSpan.content,
};
 
res.reservedTicket = {
    "@type": "Ticket",
    ticketedSeat: {
        "@type": "Seat",
        seatNumber: seatSpan.content,
    },
};
 
return res;

PKPASS

It's a object of fields inside PKPASS:

• field[X]: Object with labels and values

  Example - pkpass

  function main(pkpass, node) {
    // pass.json has "boardingPass" with keys "depar" "arrir" "arrirTime" "deparTime" "code"
    var res = node.result[0];
   
    var f = JsonLd.newFlightReservation(); // https://schema.org/FlightReservation
    f.reservationFor.departureAirport.name = pass.field["depar"].label;
    f.reservationFor.arrivalAirport.name = pass.field["arrir"].label;
    f.reservationFor.departureTime = JsonLd.toDateTime(
      pass.field["deparTime"].value,
      "hh:mm dd.MM.yyyy",
      "en",
    );
    f.reservationFor.arrivalTime = JsonLd.toDateTime(
      pass.field["arrirTime"].value,
      "hh:mm dd.MM.yyyy",
      "en",
    );
    f.reservationFor.airline.iataCode = "KD";
    f.reservationFor.flightNumber = pass.field["code"].label;
    return f; // Returns the flight reservation object later used by other apps
  }

Additional API available to extractor scripts

JSON-LD API

API for supporting schema.org output:

• JsonLd: factory functions for schema.org objects, date/time parsing, etc

More: JsonLd

Examples

var f = JsonLd.newFlightReservation(); // https://schema.org/FlightReservation
f.reservationFor.departureAirport.name = "KDE Konqi Airport (KDQ)"; // https://schema.org/FlightReservation -> https://schema.org/Flight -> https://schema.org/Place -> https://schema.org/Airport
f.reservationFor.arrivalAirport.name = "KDE Katie City Airport (KDA)";
f.reservationFor.departureTime = JsonLd.toDateTime(
    "08:36 20.02.2025",
    "hh:mm dd.MM.yyyy",
    "en",
);
f.reservationFor.arrivalTime = JsonLd.toDateTime(
    "09:56 20.02.2025",
    "hh:mm dd.MM.yyyy",
    "en",
);
f.reservationFor.airline.iataCode = "KD";
f.reservationFor.flightNumber = "KD 1096";
return f; // Returns the flight reservation object later used by other apps

ByteArray, BitArray, Barcode

API for handling specific types of input data:

• ByteArray: functions for dealing with byte-aligned binary data, including decompression, Base64 decoding, Protcol Buffer decoding, etc.
• BitArray: functions for dealing with non byte-aligned binary data, such as reading numerical data at arbitrary bit offsets. Often used if binary data is with nonstandard encoding (eg. 6bit per character).
• Barcode: functions for manual barcode decoding. This should be rarely needed nowadays, with the extractor engine doing this automatically and creating corresponding document nodes.

More: ByteArray, BitArray, Barcode

Examples

const KonqiPersonality = ByteArray.toBase64("Cheerful"); // "Q2hlZXJmdWwK"
const KatieMessage = ByteArray.fromBase64("UmVtZW1iZXIgdG8gdGFrZSBicmVha3MK"); // "Remember to take breaks"
 
const theQR = node.childNodes[1].childNodes[0].content; // Base64 encoded data
const decodedQR = ByteArray.fromBase64(theQR); // binary blob
const bitsOfQR = ByteArray.toBitArray(theQR); // Conver this to bitArray so it can be manipulated bit-by-bit
let outputString = "";
for (let i = 0; i < 6; ++i) {
    let magicalNumber = bitsOfQR.readNumberMSB(0, 6); // Reads 6 **bits**, eg. '43'
    outputString += String.fromCharCode(magicalNumber + 32); // '43' + 32 = K
}
console.log(outputString); // Konqi
 
// Usually not needed, as the extractor engine will create barcode nodes automatically
const QRCode = ImageOfAztecQRCodeNotDecodedByExtractorEngine;
const DecodedAztec = Barcode.decodeAztec(
    ImageOfAztecQRCodeNotDecodedByExtractorEngine,
);
console.log(DecodedAztec); // ["KDE airlines", "KDE Konqi Airport (KDQ)", "KDE Katie City Airport (KDA)", "20.02.2025", "08:36", "20.02.2025", "09:56", "KD 1096", "magicalstringsoweknowthisticketwasnottamperedwithbyevilwizards"]

Extractor API

API for interacting with the extractor engine itself:

• ExtractorEngine: Allows to recursively perform extraction. It can be useful for elements that need custom decoding in an extractor script first, but that contain otherwise generally supported data formats. Standard barcodes encoded in URL arguments are such an example.

More: ExtractorEngine

Examples

const XMLdataIncorreclyInterpretedAsText = "<xml><data>42</data></xml>";
const CorrectlyInterpretedXML = ExtractorEngine.extract(
    XMLdataIncorreclyInterpretedAsText,
    "application/xml",
);
 
var f = JsonLd.newFlightReservation();
ExtractorEngine.extractPrice("13 EUR", f); // Adds to ticket price

Extractor scripts

The script entry point is called with three arguments:

• The first argument is the content of the node that is processed. The data type of that argument depends on the node type as described in the document model section above. This is usually what extractor script are most concerned with.
• The second argument is the document node being processed (KItinerary::ExtractorDocumentNode, see example under). It can be useful to access already extracted results on a node (e.g. coming from generic extraction) in order to augment those.
• The third argument is the document node that matched the filter. This can be the same as the second argument (for filters with scope = Current), but it doesn't have to be. It is most useful when triggering on descendant nodes such as barcodes, the content of which will then be incorporated into the extraction result by the script.

Output of your JS function should be:

• A JS object following the schema.org ontology (JsonLd) with a single extraction result.
• A JS array containing one or more schema.org/JsonLd objects. Useful if a ticket document has multiple tickets.

Script errors and empty array is considered as "[]" (aka. nothing was returned).

Examples

Let's assume we want to create an extractor script for a railway ticket which comes with a simple tabular layout for a single leg per page, and contains a QR code with a 10 digit number for each leg.

Konqi -> Katie West
Departure: 21 Jun 18:42
Arrival: 21 Jun 23:12
 
[Big QR code]

As a filter we'd use something similar as example 2 above, triggering on the barcode content.

function extractTicket(pdf, node, barcode) {
    // text for the PDF page containing the barcode that triggered this
    const text = pdf.pages[barcode.location].text;
 
    // empty http://schema.org/TrainReservation object for the result
    let res = JsonLd.newTrainReservation();
 
    // when using regular expressions, matching on things that don't change in different
    // language variants is usually preferable, but might not always be possible
    // when creating regular expressions consider that various special characters might occur in names
    // of people or locations (in the above example spaces and parenthesis)
    const leg = text.match(/(.*) -> (.*)/); // ["Konqi", "Katie West"]
 
    // this can throw an error if the regular expression didn't match
    // that's fine though, the script is aborted here and considered not to have any result
    // ie. handling this case explicitly is unnecessary here
    res.reservationFor.departureStation.name = leg[1]; // Konqi
    res.reservationFor.arrivalStation.name = leg[2]; // Katie West
 
    // date/time parsing can recover missing year numbers from context, if available
    // In our example it would consider the PDF creation time for that, and the resulting
    // date would be the first occurrence of the given day and month following that.
    // https://doc.qt.io/qt-6/qdate.html#fromString-1
    res.reservationFor.departureTime = JsonLd.toDateTime(
        text.match(/Departure: (.*)/)[1],
        "dd MMM hh:mm",
        "en",
    );
 
    // for supporting different language formats, both the format string and the locale
    // argument can be lists. All combinations are then tried until one yields a valid result.
    res.reservationFor.arrivalTime = JsonLd.toDateTime(
        text.match(/(?:Arrival|Arrivé|Ankunft): (.*)/)[1],
        ["dd MMM hh:mm", "dd MMM hh.mm"],
        ["en", "fr", "de"],
    );
 
    // the node that triggered this script (the barcode) can be accessed and integrated into the result
    res.reservedTicket.ticketToken = "qrCode:" + barcode.content;
 
    return res;
}

The above example produces and entirely new result. Another common case are scripts that merely augment an existing result. Let's assume an Apple Wallet pass for a flight, the automatically extracted result is correct but misses the boarding group. The filter for this would be similar to example 4 above, triggering on the pass issuer.

// unused arguments can be omitted
function extractBoardingPass(pass, node) {
    // use the existing result as a starting point
    // generally this can be more than one, but specific types of documents
    // might only produce a deterministic amount (like 1 in this case).
    let res = node.result[0];
 
    // modify the result as necessary
    res.boardingGroup = pass.field["group"].label;
 
    // returning a result here will replace the existing results for this node
    return res;
}