Reorganizing and adding OVOS configs.

2025-04-22 14:02:45 -04:00
parent 10809204df
commit 0903ca039c
11 changed files with 67 additions and 102 deletions
@@ -0,0 +1,23 @@
+services:
+  beszel-agent:
+    image: henrygd/beszel-agent
+    container_name: beszel-agent
+    restart: unless-stopped
+    network_mode: host
+    volumes:
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+      # monitor other disks / partitions by mounting a folder in /extra-filesystems
+      # - /mnt/disk/.beszel:/extra-filesystems/sda1:ro
+    environment:
+      LISTEN: 45876
+      KEY: ${BEZEL_AGENT_KEY}
+  portainer-agent:
+    container_name: portainer_agent
+    image: portainer/agent:2.27.0
+    volumes:
+      - /:/host
+      - /var/lib/docker/volumes:/var/lib/docker/volumes
+      - /var/run/docker.sock:/var/run/docker.sock
+    restart: always
+    ports:
+      - 9001:9001
@@ -42,12 +42,22 @@
  // Date format, either 'MDY' (e.g. "11-29-1978") or 'DMY' (e.g. "29-11-1978")
  "date_format": "MDY",

-  // Whether to opt in to data collection
-  "opt_in": false,
-
  // Play a beep when system begins to listen?
  "confirm_listening": true,

+  // opendata servers can be hosted by users or OVOS team, either for analyzing your usage patterns or to contribute to a open dataset
+  // any number of server urls can be added here, this is EXCLUSIVELY OPT-IN
+  // source: https://github.com/OpenVoiceOS/ovos-opendata-server
+  "open_data": {
+    // NOTE: consider enabling tigregotico.pt servers to help collecting datasets!
+    // this is a server hosted by @JarbasAi and will directly help improving the intent pipeline
+    // a dashboard is provided at https://opendata.tigregotico.pt
+    // more test data is needed to complement https://gitlocalize-bench.tigregotico.pt and to identity areas of improvement
+    "intent_urls": [
+       // "https://metrics.tigregotico.pt/intents"
+    ]
+  },
+
  // File locations of sounds to play for system events
  "sounds": {
    "start_listening": "snd/start_listening.wav",
@@ -105,20 +115,6 @@
  // default to $XDG_DATA_DIRS/$BASE_FOLDER where BASE_FOLDER is read from ovos.conf (default "mycroft")
  // "cache_path": "/tmp/mycroft/cache",

-  # emit mycroft.ready signal when all these conditions are met
-  # different setups will have different needs
-  # eg, a server does not care about audio
-  # pairing -> device is paired
-  # internet -> device is connected to the internet - NOT IMPLEMENTED
-  # skills -> skills reported ready
-  # speech -> stt reported ready
-  # audio -> audio playback reported ready
-  # gui -> gui websocket reported ready - NOT IMPLEMENTED
-  # enclosure -> enclosure/HAL reported ready - NOT IMPLEMENTED
-  # network_skills -> skills with network requirements
-  # internet_skills -> skills with internet requirements
-  "ready_settings": ["skills"],
-
  // To enable a utterance transformer plugin just add it's name with any relevant config
  // these plugins can mutate the utterance between STT and the Intent stage
  // they may also modify message.context with metadata
@@ -158,6 +154,19 @@
      "conf_med": 0.8,
      "conf_low": 0.5,

+      // use snowball stemmer, helps with grammatical gender and verb tenses
+      "stem": false,
+
+      // normalization step to remove punctuation and punctuation
+      "cast_to_ascii": false,
+
+      // padaos is an internal regex matcher that ensures exact matches
+      "disable_padaos": false,
+
+      // domain engine will classify intents first and then the intent for that skill
+      // if false intents are classified directly
+      "domain_engine": false,
+      
      // path to save trained intent files
      // by default uses XDG directories if not set
      //"intent_cache": "~/.local/share/mycroft/intent_cache/",
@@ -217,7 +226,6 @@
        "fallback_high",
        "stop_medium",
        "adapt_medium",
-        "padatious_medium",
        "adapt_low",
        "common_qa",
        "fallback_medium",
@@ -231,25 +239,6 @@
    // relative to "data_dir"
    "directory": "skills",

-    // used by selene for marketplace integration in web interface
-    // NOTE: selene seems to work fine without this data
-    // no need to advertise installed skills if they dont have settings
-    // this data was tightly coupled to msm and is now mostly useless for selene
-    "upload_skill_manifest": false,
-
-    // if "sync_skill_settings" is enabled in "server" section
-    // should skill settingg changes on device be upload to selene?
-    // NOTE: this should be true, but it was removed (bug ?) in selene at some point
-    // old endpoints are however still available so functionality has been restored here
-    // it is only configurable in case you want to retain "old" mycroft-core behaviour
-    // or in case selene deprecates the old endpoint
-    "sync2way": true,
-    // values in skill settings missing in settingsmeta wont show up in selene
-    // this flag auto generates settingsmeta on the fly
-    // TODO - settings this to true may run into permission issues until settingsmeta.json gets XDG support
-    // it writes to skill base folder
-    "autogen_meta": false,
-
    // blacklisted skills to not load
    // NB: This is skill_id of the skill, usually defined in the skills setup.py
    "blacklisted_skills": [
@@ -340,78 +329,20 @@
    // eg. {"a": {"b": True, "c": False}}
    // to protect "c" you would enter "a:c" in the section below
    "protected_keys": {
-        // NOTE: selene backend expects "opt_in" to be changeable in their web ui
-        // that effectively gives them a means to enable spying without your input
-        // Mycroft AI can be trusted, but you dont need to anymore!
-        // The other keys are not currently populated by the remote backend
-        // they are defined for protection against bugs and for future proofing
-        // (what if facebook buys mycroft tomorrow?)
        "remote": [
-            "enclosure",
-            "server",
            "system",
            "websocket",
            "gui_websocket",
            "network_tests",
-            "listener:wake_word_upload:disable",
            // NOTE: selene returns listener settings as part of ww config
            // they are protected because selene has no clue about your mic setup
            "listener:channels",
-            "listener:sample_rate",
-            "listener:multiplier",
-            "listener:energy_ratio",
-            "skills:upload_skill_manifest",
-            "skills:auto_update",
-            "skills:priority_skills",
-            "skills:blacklisted_skills",
-            // NOTE: selene exposes this in web_ui, this has been disabled
-            // if you unprotect this key selene can enable opt_in behind your back
-            "opt_in"
+            "listener:sample_rate"
        ],
        "user": []
    }
  },

-  // Address of the REMOTE server
-  // Needs to be explicitly enabled, also see "protected_keys" under "system" above
-  // Possible backends
-  // - https://github.com/OpenVoiceOS/OVOS-local-backend
-  "server": {
-    // Valid types: offline, personal
-    "backend_type": "offline",
-    // url for self hosted personal backend
-    "url": "",
-    "version": "v1",
-    "update": false,
-    "metrics": true,
-    "sync_skill_settings": true
-  },
-
-  // This section controls what providers should be used by each 3rd party API
-  "microservices": {
-      // auto == backend from "server" section above
-      // auto / wolfram / personal
-      "wolfram_provider": "auto",
-      // auto / owm / personal
-      "weather_provider": "auto",
-      // auto / osm / personal
-      "geolocation_provider": "auto",
-
-      // secret keys for offline usage
-      "wolfram_key": "",
-      "owm_key": "",
-      "email": {
-          // by default send emails here
-          "recipient": "",
-          "smtp": {
-            "username": "",
-            "password": "",
-            "host": "smtp.mailprovider.com",
-            "port": 465
-          }
-      }
-  },
-
  // The ovos-core messagebus websocket
  "websocket": {
    "host": "127.0.0.1",
@@ -450,7 +381,7 @@
    "sample_rate": 16000,

    // mute global audio output volume while microphone is recording
-    "fake_barge_in": false,
+    "fake_barge_in": true,

    // TODO - these names are confusing, update dinkum listener to give them more descriptive names
    // min speech seconds for audio to be considered speech
@@ -484,7 +415,7 @@
    // "{uuid4}" - a random uuid4
    // "{now:%Y-%m-%dT%H%M%S%z}" - the local ISO datetime
    // "{utcnow:%Y-%m-%dT%H%M%S%z}" - the UTC ISO datetime
-    "utterance_filename": "{hash}-{uuid4}",
+    "utterance_filename": "{md5}-{uuid4}",

    "wake_word_upload": {
      "disable": true,
@@ -669,7 +600,13 @@
    "generic": {
        // enable/disable homescreen
        "homescreen_supported": true
-    }
+    },
+
+    // in headless devices you can set this flag to not emit any GUI related bus messages
+    // this will reduce messagebus usage
+    // NOTE: this may be undesired, GUI clients can also be cli based and
+    // do not necessarily require a desktop environment or to be running 24/7
+    "disable_gui": false
  },

  // Level of logs to store, one of  "CRITICAL", "ERROR", "WARNING", "INFO", "DEBUG"
@@ -749,9 +686,9 @@
     "module": "ovos-classifiers-postag-plugin"
  },
  // turn utterances into phoneme sequences, used to generate mouth movements
+  //  disabled by default, only relevant in devices such as the Mark 1
  "g2p": {
-     "module": "ovos-g2p-plugin-mimic",
-     "ovos-g2p-plugin-mimic": {"fallback_module": "ovos-g2p-plugin-heuristic-arpa"}
+     "module": ""
  },

  // DEPRECATED: this section is in the process of being fully removed,
@@ -836,4 +773,4 @@
  },

  "debug": false
-}
+}
@@ -0,0 +1 @@
+../ovos-docker-upstream/compose
@@ -29,6 +29,10 @@
      "ovos-PHAL-plugin-system": {
        "enabled": true
      }
+    },
+    "ovos-PHAL-plugin-dotstar": {
+      "enabled": true,
+      "dotstar_hat": "ADAFRUIT2MIC"
    }
  },
  "open_data": {