deploy: 1372c8a

_sources/melodic_analysis/pitch-extraction.ipynb

@@ -239,21 +239,54 @@
    "cell_type": "code",
    "execution_count": null,
    "metadata": {
-    "tags": []
+    "tags": [
+     "remove-output"
+    ]
    },
    "outputs": [],
    "source": [
     "# Initializing an FTANet instance\n",
     "ftanet_carnatic = compiam.load_model(\"melody:ftanet-carnatic\")\n",
     "\n",
+    "# Importing and initializing again a melodia instance of comparison\n",
+    "from compiam.melody.pitch_extraction import Melodia\n",
+    "melodia = Melodia()  "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "We will now initialize Saraga Carnatic dataset to get an example track, load it, and run inference using FTANet-Carnatic and Melodia, both accessed through `compiam`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "tags": [
+     "remove-output"
+    ]
+   },
+   "outputs": [],
+   "source": [
+    "import librosa\n",
+    "from compiam import load_dataset\n",
+    "\n",
+    "saraga_carnatic = load_dataset(\n",
+    "    \"saraga_carnatic\",\n",
+    "    data_home=os.path.join(\"..\", \"audio\", \"mir_datasets\")\n",
+    ")\n",
+    "saraga_tracks = saraga_carnatic.load_tracks()\n",
+    "example = saraga_tracks[\"109_Sri_Raghuvara_Sugunaalaya\"]\n",
+    "\n",
+    "# Loading the audio\n",
+    "y, sr = librosa.load(example.audio_path)\n",
+    "y = y[:3*sr]  # Getting just the first 3 seconds\n",
+    "\n",
     "# Predict!\n",
-    "ftanet_pitch_track = ftanet_carnatic.predict(\n",
-    "    os.path.join(\n",
-    "        \"..\", \"audio\", \"mir_datasets\", \"saraga1.5_carnatic\",\n",
-    "        \"Live at Vani Mahal by Sanjay Subrahmanyan\", \"Sri Raghuvara Sugunaalaya\",\n",
-    "        \"Sanjay Subrahmanyan - Sri Raghuvara Sugunaalaya.mp3\"\n",
-    "    ),\n",
-    ")"
+    "ftanet_pitch_track = ftanet_carnatic.predict(example.audio_path)\n",
+    "melodia_pitch_track = melodia.extract(example.audio_path)"
    ]
   },
   {
@@ -274,15 +307,6 @@
     "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
     "\n",
-    "y, sr = librosa.load(\n",
-    "    os.path.join(\n",
-    "        \"..\", \"audio\", \"mir_datasets\", \"saraga1.5_carnatic\",\n",
-    "        \"Live at Vani Mahal by Sanjay Subrahmanyan\", \"Sri Raghuvara Sugunaalaya\",\n",
-    "        \"Sanjay Subrahmanyan - Sri Raghuvara Sugunaalaya.mp3\"\n",
-    "    )\n",
-    ")\n",
-    "y = y[:3*sr]  # Getting just the first 3 seconds for better visualizatiom\n",
-    "\n",
     "fig, ax = plt.subplots(nrows=1, ncols=1, sharex=True, figsize=(15, 12))\n",
     "D = librosa.amplitude_to_db(np.abs(librosa.stft(y)), ref=np.max)\n",
     "img = librosa.display.specshow(D, y_axis='linear', x_axis='time', sr=sr, ax=ax);\n",

_sources/melodic_analysis/raga-recognition.ipynb

@@ -93,6 +93,15 @@
     "deepsrgm = compiam.load_model(\"melody:deepsrgm\")"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "```{important}\n",
+    "Deep Learning model checkpoints tend to be large in size, therefore storing these in `compiam` may become unsustainable. For that reason, we store the checkpoints in the cloud and download these when the user initializes a model using the wrapper. Note that you can specify to which location the checkpoint should be donwloaded by specifying `data_home` argument in `load_model()`.\n",
+    "```"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -179,7 +188,7 @@
     "deepsrgm.predict(feat)\n",
     "```\n",
     "\n",
-    "**Oops! Why is this line of code not running?** This ``.predict()`` method runs inference with the DEEPSRGM model using the passed features. As you may now, DL models tend to occupy an important amount of memory from your computer to run, especially in the training stage. Some light-weight models can be run from a conventional laptop, but in some other cases you might need a machine with enough power to run your models and perform inference. \n",
+    "**Oops! Why is this line of code not being executed?** This ``.predict()`` method runs inference with the DEEPSRGM model using the passed features. As you may now, DL models tend to occupy an important amount of memory from your computer to run, especially in the training stage. Some light-weight models can be run from a conventional laptop, but in some other cases you might need a machine with enough power to run your models and perform inference. \n",
     "\n",
     "```{tip}\n",
     "As mentioned in the [introduction](google-collab), Google Collab has GPU and TPU access which may allow you running certain models that are too big for your machine.\n",

_sources/resources/citing.md

@@ -22,7 +22,7 @@ The cite the book as a whole please cite:
 We now explicitly attribute each author to their contribution, in case you need to cite bits of text or code instead of the book as a while.
 * [Setup](welcome-setup), [Python](python) and [compIAM](compiam), [corpora](corpora) and [datasets](datasets), and tool walkthroughs: **Thomas Nuttall, Genís Plaja-Roglans, Lara Pearson and Brindha Manickavasakan, Xavier Serra**
 * [What is Indian Art Music?](indian-art-music) and introduction to Carnatic Music ([instrumentation](carnatic-instrumentation), [format](carnatic-formats), and [melodic concepts](carnatic-melodic-concepts)): **Lara Pearson and Brindha Manickavasakan**
-* Introduction to Hindustani Music ([instrumentation](hindustani-instrumentation), [format](hindustani-formats), and [melodic concepts](hindustani-melodic-concepts)): **Kaustuv Kanti Ganguli**
+* [Introduction to Hindustani Music](hindustani-music): **Kaustuv Kanti Ganguli**
 * [Carnatic Rhythm](carnatic-rhythm) and [Hindustani Rhythm](hindustani-rhythm): **Ajay Srinivasamurthy**
 
 If you use `compiam` please cite as:

_sources/rhythmic_analysis/meter_analysis.ipynb

@@ -19,7 +19,11 @@
    },
    "outputs": [],
    "source": [
-    "## Importing compiam to the project\n",
+    "## Installing (if not) and importing compiam to the project\n",
+    "import importlib.util\n",
+    "if importlib.util.find_spec('compiam') is None:\n",
+    "    ## Bear in mind this will only run in a jupyter notebook / Collab session\n",
+    "    %pip install compiam\n",
     "import compiam\n",
     "\n",
     "# Import extras and supress warnings to keep the tutorial clean\n",
@@ -238,7 +242,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "predicted_aksharas[:20]"
+    "predicted_aksharas"
    ]
   },
   {
@@ -277,7 +281,7 @@
     "\n",
     "# And we plot!\n",
     "plot_waveform(\n",
-    "    file_path=track.audio_path,\n",
+    "    input_data=track.audio_path,\n",
     "    t1=0,\n",
     "    t2=4,\n",
     "    labels=predicted_beats_dict,\n",
@@ -293,7 +297,7 @@
   },
   "language_info": {
    "name": "python",
-   "version": "3.9.4"
+   "version": "3.11.6"
   },
   "orig_nbformat": 4,
   "vscode": {