Login

Snippets by limodou

Snippet List

Format transition middleware

Note: This is a testing middleware. This snippets may be changed frequently later. What's it ----------- Sometimes I thought thow to easy the output data into another format, except html format. One way, you can use decorator, just like: @render_template(template='xxx') def viewfunc(request,...): And the output data of viewfunc should be pure data. And if want to output json format, you should change the decorator to: @json_response def viewfunc(request,...): I think it's not difficult. But if we can make it easier? Of cause, using middleware. So you can see the code of `process_response`, it'll judge the response object first, if it's an instance of HttpResponse, then directly return it. If it's not, then get the format of requst, if it's `json` format, then use json_response() to render the result. How to setup `request.format`? In `process_request` you and see, if the `request.REQUEST` has a `format` (you can setup it in settings.py with FORMAT_STRING option), then the `request.format` will be set as it. If there is not a such key, then the default will be `json`. So in your view code, you can just return a python variable, this middleware will automatically render this python variable into json format data and return. For 0.2 it support xml-rpc. But it's very different from common implementation. For server url, you just need put the same url as the normal url, for example: http://localhost:8000/booklist/ajax_list/?format=xmlrpc Notice that the format is 'xmlrpc'. A text client program is: from xmlrpclib import ServerProxy server = ServerProxy("http://localhost:8000/booklist/ajax_list/?format=xmlrpc", verbose=True) print server.booklist({'name':'limodou'}) And the method 'booklist' of server is useless, because the url has include the really view function, so you can use any name after `server`. And for parameters of the method, you should use a dict, and this dict will automatically convert into request.POST item. For above example, `{'name':'limodou'}`, you can visit it via `request.POST['name']` . For `html` format, you can register a `format_processor` callable object in `request` object. And middleware will use this callable object if the format is `html`. Intall --------- Because the view function may return non-HttpResponse object, so this middleware should be installed at the end of MIDDLEWARE_CLASSES sections, so that the `process_response` of this middleware can be invoked at the first time before others middlewares. And I also think this mechanism can be extended later, for example support xml-rpc, template render later, etc, but I have not implemented them, just a thought. Options --------- FORMAT_STRING used for specify the key name of format variable pair in QUERY_STRING or POST data, if you don't set it in settings.py, default is 'format'. DEFAYLT_FORMAT used for default format, if you don't set it in settings.py, default is 'json'. Reference ----------- Snippets 8 [ajax protocol for data](http://www.djangosnippets.org/snippets/8/) for json_response

  • middleware
  • format
  • json
Read More

Another pygments for ReST

This is like [snippet 36](/snippets/36/). And it'll return css also. And it's not a filter.If the code parameter is skip, it'll test the code first, and if there is not a suiable lexer for the code, then use default python lexer to render the code. The code lanauage parameter is comes from pygments lexer alias. How to use it ---------------- html = to_html(rest_text) And there is a level parameter in to_html function, default is `2`, it's the sections level. And the html will be `css style + body` How to write ReST --------------------- Below is a example. This is a test. .. code:: def code(name, arguments, options, content, lineno, content_offset, block_text, state, state_machine): opt = {'display':'on'} opt.update(options) docnodes.Node(content, ''.join(arguments), **opt) if opt['display'].lower() == 'on': return [nodes.literal_block('', '\n'.join(content))] else: return [] .. code:: html+django <h1 id="title">通讯录</h1> <hr> <div> <table border="0" width="500"> <tr align="right"> <td>{% if has_previous %} <a href="/address?page={{ previous }}">上一页</a> {% endif %} {% if has_next %} <a href="/address?page={{ next }}">下一页</a> {% endif %}</td></tr> </table>

  • pygments
  • rest
Read More

Validator for data

This module is not aimed to replace the newforms, but I would like manually write html code, and just need a pure validate module, so I write this, and many things may be similar with newforms. So if you like me would only need a pure validator module, you can use it. And it has some different features from newforms: 1. Support validator_list parameter, so you could use it just like the old manipuator class 2. Supply easy method, such as `validate_and_save()`, so you can pass a request object, and get a tuple result `(flag, obj_or_error)`, if the `flag` is `True`, then the next value is an object; and if the `flag` is `False`, then the next value is error message. 3. Each field has a `validate_and_get` method, and it'll validate first and then return the result, maybe an object or error message. Just like above. 4. SplitDateTimeField is somewhat different from the newforms. For example:: c = SplitDateTimeField('date', 'time') print c.validate_and_get({'date':'2006/11/30', 'time':'12:13'}) So the first parameter is DateField's field_name, and the second parameter is TimeField's field_name. 5. Add yyyy/mm/dd date format support 6. Support default value of a field. You can add a default value for a field, if this field is not required, and the value is *empty*, Validator will return the default value. This module is new, so many things could be changed.

  • validator
Read More