Write opencv frames into gstreamer rtsp server pipeline

I'm trying to put opencv images into a gstreamer rtsp server in python. I have some issue writing in the mediafactory, I'm new to gst-rtsp-server ancd there's little documentation so I don't know exactly if I'm using the right approach. I'm using a thread to start the MainLoop and I'm using the main thread to create a buffer to push in the appsrc element of the mediafactory pipeline. Am I using the right approach to obtain my objective? Can anyone help me? My code is below:

from threading import Thread
from time import clock

import cv2
import gi

gi.require_version('Gst', '1.0')
gi.require_version('GstRtspServer', '1.0')
from gi.repository import Gst, GstRtspServer, GObject


class SensorFactory(GstRtspServer.RTSPMediaFactory):
    def __init__(self, **properties):
        super(SensorFactory, self).__init__(**properties)
        self.launch_string = 'appsrc ! video/x-raw,width=320,height=240,framerate=30/1 ' \
                             '! videoconvert ! x264enc speed-preset=ultrafast tune=zerolatency ' \
                             '! rtph264pay config-interval=1 name=pay0 pt=96'
        self.pipeline = Gst.parse_launch(self.launch_string)
        self.appsrc = self.pipeline.get_child_by_index(4)

    def do_create_element(self, url):
        return self.pipeline


class GstServer(GstRtspServer.RTSPServer):
    def __init__(self, **properties):
        super(GstServer, self).__init__(**properties)
        self.factory = SensorFactory()
        self.factory.set_shared(True)
        self.get_mount_points().add_factory("/test", self.factory)
        self.attach(None)


GObject.threads_init()
Gst.init(None)

server = GstServer()

loop = GObject.MainLoop()
th = Thread(target=loop.run)
th.start()

print('Thread started')

cap = cv2.VideoCapture(0)

print(cap.isOpened())

frame_number = 0

fps = 30
duration = 1 / fps

timestamp = clock()

while cap.isOpened():
    ret, frame = cap.read()
    if ret:

        print('Writing buffer')

        data = frame.tostring()

        buf = Gst.Buffer.new_allocate(None, len(data), None)
        buf.fill(0, data)
        buf.duration = fps
        timestamp = clock() - timestamp
        buf.pts = buf.dts = int(timestamp)
        buf.offset = frame_number
        frame_number += 1
        retval = server.factory.appsrc.emit('push-buffer', buf)
        print(retval)

        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

cap.release()

By the way I tried to copy the buffer creation from opencv source code but I'm not sure I correctly trandlated the c++ code in python.

I found the solution, lots of stuff was missing.

• I used the need-data signal like the examples of gst-rtsp-server.
  
• I changed some default parameters of appsrc, like the is-live, block and format.
  
• The caps on the appsrc element.
• I was not setting the offset of the buffer correctly.

Here's the code for anyone who's facing the same problem or has a somewhat similar one.

#!/usr/bin/env python3

import cv2
import gi

gi.require_version('Gst', '1.0')
gi.require_version('GstRtspServer', '1.0')
from gi.repository import Gst, GstRtspServer, GObject


class SensorFactory(GstRtspServer.RTSPMediaFactory):
    def __init__(self, **properties):
        super(SensorFactory, self).__init__(**properties)
        self.cap = cv2.VideoCapture(0)
        self.number_frames = 0
        self.fps = 30
        self.duration = 1 / self.fps * Gst.SECOND  # duration of a frame in nanoseconds
        self.launch_string = 'appsrc name=source is-live=true block=true format=GST_FORMAT_TIME ' \
                             'caps=video/x-raw,format=BGR,width=640,height=480,framerate={}/1 ' \
                             '! videoconvert ! video/x-raw,format=I420 ' \
                             '! x264enc speed-preset=ultrafast tune=zerolatency ' \
                             '! rtph264pay config-interval=1 name=pay0 pt=96'.format(self.fps)

    def on_need_data(self, src, lenght):
        if self.cap.isOpened():
            ret, frame = self.cap.read()
            if ret:
                data = frame.tostring()
                buf = Gst.Buffer.new_allocate(None, len(data), None)
                buf.fill(0, data)
                buf.duration = self.duration
                timestamp = self.number_frames * self.duration
                buf.pts = buf.dts = int(timestamp)
                buf.offset = timestamp
                self.number_frames += 1
                retval = src.emit('push-buffer', buf)
                print('pushed buffer, frame {}, duration {} ns, durations {} s'.format(self.number_frames,
                                                                                       self.duration,
                                                                                       self.duration / Gst.SECOND))
                if retval != Gst.FlowReturn.OK:
                    print(retval)

    def do_create_element(self, url):
        return Gst.parse_launch(self.launch_string)

    def do_configure(self, rtsp_media):
        self.number_frames = 0
        appsrc = rtsp_media.get_element().get_child_by_name('source')
        appsrc.connect('need-data', self.on_need_data)


class GstServer(GstRtspServer.RTSPServer):
    def __init__(self, **properties):
        super(GstServer, self).__init__(**properties)
        self.factory = SensorFactory()
        self.factory.set_shared(True)
        self.get_mount_points().add_factory("/test", self.factory)
        self.attach(None)


GObject.threads_init()
Gst.init(None)

server = GstServer()

loop = GObject.MainLoop()
loop.run()