#!/usr/bin/env bash

# Prepare script for MLS English ASR recipe in icefall

# fix segmentation fault reported in https://github.com/k2-fsa/icefall/issues/674
export PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python

set -eou pipefail

stage=-1
stop_stage=100

# Configuration for BPE tokenizer
vocab_sizes=(2000)  # You can add more sizes like (500 1000 2000) for comparison

# Directory where dataset will be downloaded
dl_dir=$PWD/download

. shared/parse_options.sh || exit 1

# All files generated by this script are saved in "data".
mkdir -p data
mkdir -p data/audio
mkdir -p data/manifests
mkdir -p data/lang

log() {
	local fname=${BASH_SOURCE[1]##*/}
	echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*"
}

log "Starting MLS English data preparation"

if [ $stage -le 0 ] && [ $stop_stage -ge 0 ]; then
	log "Stage 0: Download MLS English dataset"
	# Check if huggingface_hub is installed
	if ! python -c "import huggingface_hub" &> /dev/null; then
		log "huggingface_hub Python library not found. Installing it now..."
		# Using --break-system-packages for Debian/Ubuntu environments where pip install might fail without it
		python -m pip install huggingface_hub || \
			python -m pip install huggingface_hub --break-system-packages || { \
			log "Failed to install huggingface_hub. Please install it manually: pip install huggingface_hub"; \
			exit 1; \
		}
		log "huggingface_hub installed successfully."
	fi

  # Check if the dataset already exists to avoid re-downloading
  if [ ! -d "$dl_dir/mls_english" ]; then
	  log "Dataset not found at $dl_dir/mls_english. Starting download..."
	  if ! python ./local/utils/download_mls_english.py --dl-dir "$dl_dir"; then
		  log "Failed to download MLS English dataset via download_mls_english.py"
		  exit 1
	  fi
  else
	  log "Dataset already exists at $dl_dir/mls_english. Skipping download."
  fi
fi

if [ $stage -le 1 ] && [ $stop_stage -ge 1 ]; then
	log "Stage 1: Compute MLS English fbank"
	if [ ! -e data/manifests/.mls_english-validated.done ]; then
		python local/compute_fbank_mls_english.py \
			--manifest-dir data/manifests \
			--audio-dir data/audio \
			--dl-dir $dl_dir/mls_english
					# --dl-dir /root/datasets/parler-tts--mls_eng
					python local/validate_manifest.py --manifest data/manifests/mls_eng_cuts_train.jsonl.gz
					python local/validate_manifest.py --manifest data/manifests/mls_eng_cuts_dev.jsonl.gz
					python local/validate_manifest.py --manifest data/manifests/mls_eng_cuts_test.jsonl.gz
					touch data/manifests/.mls_english-validated.done
	fi
fi

if [ $stage -le 2 ] && [ $stop_stage -ge 2 ]; then
	log "Stage 2: Prepare transcript for BPE training"
	if [ ! -f data/lang/transcript.txt ]; then
		log "Generating transcripts for BPE training"
		python local/utils/generate_transcript.py \
			--dataset-path $dl_dir/mls_english \
			--lang-dir data/lang \
			--split train 
	fi
fi

if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
	log "Stage 3: Prepare BPE tokenizer"
	for vocab_size in ${vocab_sizes[@]}; do
		log "Training BPE model with vocab_size=${vocab_size}"
		bpe_dir=data/lang/bpe_${vocab_size}
		mkdir -p $bpe_dir

		if [ ! -f $bpe_dir/bpe.model ]; then
			python local/train_bpe_model.py \
				--lang-dir $bpe_dir \
				--vocab-size $vocab_size \
				--transcript data/lang/transcript.txt
		fi
	done
fi

if [ $stage -le 4 ] && [ $stop_stage -ge 4 ]; then
	log "Stage 4: Show manifest statistics"
	python local/display_manifest_statistics.py --manifest-dir data/manifests > data/manifests/manifest_statistics.txt
	cat data/manifests/manifest_statistics.txt
fi

log "MLS English data preparation completed successfully"