not able to record call

[newpond]

hi I have a menu automation which calls the recording script as noted in the configuration but cannot get it to record a file. Can the file name be created on the fly rather using the same one all the time?

do you have any examples of how to run the recording function?

[arnonym]

@dirtyharryiv Can you take a look here?

[dirtyharryiv]

Yes, you can choose any file name you want. However, it must be an absolute path that the add-on has permission to write to. This should be anything below your /config/ and /media/ folders.
I use Pyscript for most of my automations, so unfortunately I don't have any YAML examples. Essentially, you answer the call with the answer command, and then send the record command. If you would like, I can provide a Pyscript example.
If a file is created but there is no sound, bear in mind that HA-SIP is only recording the remote audio not the local media or tts.

[arnonym]

For changing the file name every time, you can use home assistant template function:

service: hassio.addon_stdin
data:
  addon: c7744bff_ha-sip
  input:
    command: start_recording
    number: sip:**620@fritz.box
    recording_file: >
      /config/www/call_{{ now().strftime('%Y%m%d_%H%M%S') }}.wav

Untested, but you should get the idea.

[newpond]

For changing the file name every time, you can use home assistant template function:

service: hassio.addon_stdin
data:
addon: c7744bff_ha-sip
input:
command: start_recording
number: sip:**620@fritz.box
recording_file: >
/config/www/call_{{ now().strftime('%Y%m%d_%H%M%S') }}.wav
Untested, but you should get the idea.

yes this is what I was in need of, I will test that when I get the recording itself working!!! thanks Arne

[newpond]

Yes, you can choose any file name you want. However, it must be an absolute path that the add-on has permission to write to. This should be anything below your /config/ and /media/ folders. I use Pyscript for most of my automations, so unfortunately I don't have any YAML examples. Essentially, you answer the call with the answer command, and then send the record command. If you would like, I can provide a Pyscript example. If a file is created but there is no sound, bear in mind that HA-SIP is only recording the remote audio not the local media or tts.

HI , thanks for your swift reply. What I am doing is using an automation to calls a user that has the linphone client installed (either on mobile or tablet). So im not using the ha-sip answer command to answer the call. does this make a difference? I am calling the record command from the originator (my HA system)

my script currently transfers the call from the ha system itself to a linphone app running on the host machine (which automatically answers the call and starts recording it) I want to avoid this though and record it directly through ha-sip though as teh method I use now means I have to create two sip accounts to get it to work, plus just basically a messy solution. (and the call doesnt always terminate properly after hang up)

I would be grateful if you could send your pyscript as mentioned as I dont use these so interested in seeing the syntax.

addon: c7744bff_ha-sip
input:
  command: dial
  number: sip:n***********
  cache_audio: true
  wait_for_audio_to_finish: true
  sip_account: 2
  menu:
    message: >-
      Please press 1 to leave a voice message now, or press 2 to be called back
      in 2 hours time
    choices:
      "1":
        id: record
        message: >-
          Please speak slowly and clearly, you may hangup when you have
          finished.
        action:
          domain: script
          service: turn_on
          entity_id: script.call_transfer
        post_action: hangup
      "2":
        id: later
        message: Ok, i will call you back later
        post_action: hangup
      default:
        id: wrong_code
        message: Wrong code, please try again
        post_action: return
      timeout:
        id: timeout
        message: Bye.
        post_action: hangup

cheers
Jon

[newpond]

oh and incidentally I dont have any tts or stt entries in my configuration.yaml as using wyoming-microsoft-stt and wyoming-microsoft-tts add-on's which doesn't (seem) to require them as they work fine (unless that is what Im missing!!

[dirtyharryiv]

So im not using the ha-sip answer command to answer the call. does this make a difference?

No, that should not make any difference. Just be sure the call is established when you send the record command. You could use the call_established webhook for this.

Here is something chatgpt came up with that looks promising but untested:

alias: Call Established – Start Recording
description: Startet die Aufnahme über Addon stdin nach Webhook
mode: single

trigger:
  - platform: webhook
    webhook_id: ha-call-established

action:
  - service: hassio.addon_stdin
    data:
      addon: c7744bff_ha-sip
      input:
        command: start_recording
        number: <the_number_you_were_calling>
        recording_file: >
          /config/www/call_{{ now().strftime('%Y%m%d_%H%M%S') }}.wav

I use ha-sip as an answering machine. So i am not placing a call but i receive one. Here is a pyscript example for that:

from datetime import datetime

VOICE = "de_DE-Homer-high"
ADDON_ID = "c7744bff_ha-sip"
RECORDING_DIR = "/media/ha_media/call_recordings/"

menu_answering_machine = {"id": "main",
                    "post_action": "noop",
                    "message":"Hello there! Your message is being recorded.",
                    "voice":VOICE,
                    "cache_audio": "true"
                }

@webhook_trigger("name_of_the_webhook_in_ha-sip")
def somebodyIsCalling(payload):
    event = payload["event"]
    caller = payload["caller"]
    parsed_caller = payload["parsed_caller"]

    if event == "incoming_call":
        log.info(f"Somebody is calling! {caller} ({parsed_caller})")
        service.call("notify", "mobile_phone_entity", message=f"You are being called by {caller}!")
        service.call("hassio", "addon_stdin", addon=ADDON_ID, input={"command":"answer", "number":parsed_caller, "menu":menu_answering_machine})
        service.call("hassio", "addon_stdin", addon=ADDON_ID, input={"command":"start_recording", "number":parsed_caller, "recording_file":getRecordingFilename()})

    elif event == "recording_stopped":
        recording_file_ret = payload["recording_file"]
        log.info(f"New voice record! File: {recording_file_ret} caller {parsed_caller}")
        task.sleep(5) # wait for file being written
        # Send recorded message via telegram
        service.call("telegram_bot", "send_voice", target="telegram_chat_id", caption=f"Audio von: {parsed_caller}!", file=recording_file_ret, config_entry_id="config_entry_id")

def getRecordingFilename():
    timestamp = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
    recording_file = RECORDING_DIR + timestamp + ".wav"
    return recording_file

I hope i could help.

Cheers Harry

[dirtyharryiv]

If i understand you correct this is kind of what you want:

alias: SIP Call – Wait for Established – Start Recording
description: Calls via ha-sip, waits for webhook "entered_menu", starts recording.
mode: single

trigger:
  - platform: event
    event_type: call_sip_recording_start

variables:
  sip_number: "sip:**622@192.168.0.1"
  recording_path: "/config/www/call_{{ now().strftime('%Y%m%d_%H%M%S') }}.wav"

action:
  - service: hassio.addon_stdin
    data:
      addon: c7744bff_ha-sip
      input:
        command: dial
        number: "{{ sip_number }}"
        cache_audio: true
        wait_for_audio_to_finish: true
        sip_account: 2
        webhook_to_call:
          entered_menu: ha-entered-menu
        menu:
          message: >-
            Please press 1 to leave a voice message now, or press 2 to be called back
            in 2 hours time
          choices:
            "1":
              id: record
              message: >-
                Please speak slowly and clearly, you may hangup when you have finished.
              post_action: noop
            "2":
              id: later
              message: Ok, i will call you back later
              post_action: hangup
            default:
              id: wrong_code
              message: Wrong code, please try again
              post_action: return
            timeout:
              id: timeout
              message: Bye.
              post_action: hangup

  - wait_for_trigger:
      - platform: webhook
        webhook_id: ha-entered-menu
    timeout: "00:01:30"
    continue_on_timeout: false
    
  - condition: template
    value_template: >
      {{ wait.trigger.json.menu_id == 'record' }}

  - service: hassio.addon_stdin
    data:
      addon: c7744bff_ha-sip
      input:
        command: start_recording
        number: "{{ sip_number }}"
        recording_file: "{{ recording_path }}"

Tested by manually triggering. Be sure to change the number and path variables and probably the trigger.
Edit: Removed unnecessary code

[newpond]

thanks for your help here, yes that looks pretty good!! however the bit that is confusing me is
choices:
"1":
id: record
message: >-
Please speak slowly and clearly, you may hangup when you have finished.
action:
domain: script
service: turn_on
entity_id: script.call_transfer
post_action: hangup

this was the action I used to transfers the call to linphone client to record the calls, so the bit I want to avoid.

what action should I use here instead? and would the call and the recording stop when the recipient hangs up?

previously I was using a script for this but presume this is an automation?

presumably I can just change the trigger to start it when a device is triggered fro example?

thanks in advance for all your help!!
cheers
Jon

[dirtyharryiv]

I updated my comment.

Yes , this is an automation. I hardly use scripts and automations, always thought they have the same syntax.
You should be able to set the trigger to whatever you want (device, time, etc.).

thanks for your help here, yes that looks pretty good!! however the bit that is confusing me is
choices:
"1":
id: record
message: >-
Please speak slowly and clearly, you may hangup when you have finished.
action: domain: script service: turn_on entity_id: script.call_transfer post_action: hangup

You do not need that part anymore. You can remove that whole action part and just leave post_action: noop. As soon as you press "1" you get inside the "record" menu, the webhook triggers and the menu id is checked for record. Then recording should start.

The recording will end automatically when the call has ended.

Cheers
Harry

[newpond]

hi Harry and Arne

weirdly it wouldnt send out the message or record anything, but all of a sudden it started working!!! and is recording the message too.
so all looking good!! I'm going to do a bit more testing and will get back to you and let you know if everything goes ok.
thank you so much for you help I really appreciate it.

thanks guys

cheers
Jon

UPDATE it turns out it wasnt triggering because the trigger button wasnt switching on the automation!!!

[newpond]

Hi Guys

one thing I have noticed with this automation is that it seems to start recording as soon as it gets the web hook saying option 1 has been selected which is before the message saying speak clearly etc has finished being announced. This means that when the recording file is played back, there is a few seconds (the length of the speak clearly announcement) of white noise at the start.
I've tried adding a time delay to the automation but it doesnt seem to make any difference . (if anything it seems to add to the length of the white noise period.

I also was hoping to add this to the end of the speak clearly announcement if possible, but it seems to cut in before the announcement had finished...any thoughts?

addon: c7744bff_ha-sip
input:
command: play_audio_file
number: "{{ sip_number }}"
audio_file: /config/www/Sound_Files/beep_1.mp3
cache_audio: false
wait_for_audio_to_finish: true
post_action: noop

cheers
Jon

[dirtyharryiv]

You could try to wait for the webhook playback_done after entered_menu.

[newpond]

hi harry, sorry I dont quite understand do you mean after the web hook payload has been received?
I tried it after the web Hook trigger section but before the call recording section and that didn't work, even with a four second delay added. (which the call recording script seem to ignore as it started recording before the 4 second delay had completed for some reason?) do then not run in the order they are listed in if there is a web hook involved?