Create another script called upload.py:
def index(): form = FORM(action="upload",method="post",enctype="multipart/form-data") form <= INPUT(Type="file",name="src") form <= INPUT(Type="submit",value="Ok") return HTML(BODY(form)) def upload(src): import mimetypes gtype,encoding = mimetypes.guess_type(src.filename) RESPONSE_HEADERS['Content-type'] = gtype data = src.file.read() return data
The function index() looks like the previous example. The changes are the `enctype` parameter of the FORM, set to "multipart/form-data"; and the type of the first INPUT field.
In the upload function, we receive the argument `src`. Since it was sent as a file inside a multipart/form-data form, this object has an attribute filename matching the name of the uploaded file ; it also has an attribute file, a file-like object with a method read() that returns bytes.
The Python module mimetypes can guess the type of this file by its extension ; for instance if the uploaded file was a JPEG image, with the extension `.jpg`, the result of mimetypes.guess_type(src.filename) will be "image/pjpeg".
RESPONSE_HEADERS['Content-type'] = gtype
introduces another name which is always available in Python scripts run by Karrigell :
RESPONSE_HEADERS. It is a dictionary-like object (in fact, an instance of class
email.message.Message, used to set the response headers that will be sent to the web browser. Here, the line sets the Content-type header to the type of the file received.
In the next line we read the binary data from the attribute "file" of src, and finally return this binary data. The Karrigell engine sends the response headers, then the binary data ; with this information, the browser renders the file content under the appropriate format : if the file was a JPEG picture, the browser displays it.