srt2txt

Convert SRT subtitle files into clean plain text. Fast & parallel with safe defaults and composable post‑processing.

Features

• Removes numbering & timestamps
• Strips simple HTML tags (<i>, <b>, etc.)
• Joins multi‑line caption blocks into single logical lines
• Join multiple files into one (--join) OR flatten all captions into one sentence stream (--join-sentences)
• Parallel file processing via Rayon
• Optional immediate duplicate removal & blank line collapsing
• Refuses overwrites unless --force

Install

cargo install srt2txt

Quick Start

# Basic conversion (creates input.txt next to input.srt)
srt2txt input.srt

# Convert a directory tree recursively → outputs into cleaned/
srt2txt subtitles/ --output-dir cleaned/

# Join several files into combined.txt (default name)
srt2txt a.srt b.srt --join

# Custom join filename
srt2txt a.srt b.srt --join --join-name all_dialogue.txt

# Stream per‑file results to stdout (no files written)
srt2txt movie.srt --stdout

# Single continuous text block (no blank separators)
srt2txt movie.srt --join-sentences --stdout

Sample Input / Output

Input (sample.srt):

1
00:00:00,000 --> 00:00:01,000
<i>Hello</i>  world!

2
00:00:01,500 --> 00:00:03,000
Second  line.
Line continued.

Output (sample.txt):

Hello world!

Second line. Line continued.

Flags

--stdout              Print combined per-file outputs to stdout (disables file writes)
--join                Aggregate all processed files into a single output (excludes --stdout, --join-sentences)
--join-name <NAME>    Filename for --join output (default: combined.txt)
--join-sentences      Flatten all captions into one continuous block (ignores --collapse-blank; conflicts with --join)
--collapse-blank      Collapse ≥3 consecutive newlines to a single blank line
--remove-duplicates   Remove immediately repeated normalized lines
-f, --force           Overwrite existing output files
-o, --output-dir DIR  Directory for outputs (created if missing)

Flag Interaction Matrix

Edge Cases & Behavior

• Duplicate sequence numbers: accepted if timestamps parse.
• Malformed timestamp lines: block skipped (robust > strict).
• Non-.srt files ignored; directories walked recursively.
• HTML stripping is naive (<[^>]+>): removes any simple tag.
• Dedup removes only immediately consecutive identical cleaned lines.

Performance Notes

• File-level parallelism only (predictable memory & ordering).
• Regex patterns compiled once via lazy_static.
• Per-file processing builds a vector of cleaned caption strings; large joined output materialized only at final aggregation.

Development

Core logic resides in src/main.rs (parsing, cleaning, post-process, CLI validation).

Build & test locally:

cargo build
cargo test --all --quiet
cargo fmt --all -- --check
cargo clippy --all-targets --all-features -- -D warnings

Git Hooks (cargo-husky)

Installed/updated on cargo build according to Cargo.toml:

Skip temporarily:

HUSKY=0 git commit -m "wip"

Force regenerate if stale:

cargo clean -p cargo-husky || true
cargo build

Contributing

Keep features lean. Add unit tests beside changed logic (see existing tests in src/main.rs). Maintain flag validation in main. When modularizing, prefer pure helpers (e.g. parser.rs, transform.rs) while preserving existing behavior.

License

MIT