project update from 24Z WIMU

24Z_update.md

@@ -0,0 +1,81 @@
+## Added functionalities:
+
+- Miditok library upgrade: from 2.1.7 to 3.0.4
+- Added PerTok tokenizator
+- New logics of corresponding notes with tokens (forced by library upgrade)
+- New parameters in configuration (mostly for PetTok): use_programs, use_microtiming, ticks_per_quarter, max_microtiming_shift, num_microtiming_bins
+- Default tokenizator classes, now associating notes to tokens is run after we receive tokens, unifed functions for all tokenizers depending on "use_programs" parameter usage 
+
+### frontend:
+
+- New design in general
+- Subtle animations
+- New piano roll- high resoultion, stable keybord so it does not vanish with scrolling
+- Tokens can be paginated and scaled
+- New audio player
+
+![Screenshot of app](img/app_screenshot.png)
+
+## Deployment
+
+You can see an example deployment on Railway [here](https://miditok-visualizer-production-frontend.up.railway.app/)
+As long as we're above our trial limit on Railway 
+
+## Running locally:
+
+You can use same commands as before to run the project locally.
+
+### Frontend
+
+Basic run:
+
+```sh
+cd frontend
+npm install
+npm run dev
+```
+
+Using Docker:
+
+```sh
+cd frontend
+docker build . -t frontend
+docker run frontend -p 3000:3000
+```
+
+### Backend
+
+Basic run:
+
+```sh
+cd backend
+poetry shell
+poetry install
+python -m core.main
+```
+
+or
+
+```sh
+poetry run python -m core.main
+```
+
+Using Docker:
+
+```sh
+cd backend
+DOCKER_BUILDKIT=1 docker build --target=runtime . -t backend
+docker run backend -p 8000:8000
+```
+
+### All at once:
+
+```sh
+docker-compose up
+```
+
+## Contributors
+
+- Karol Bogumił
+- Maksymilian Banach
+- Jakub Przesmycki

README.md

@@ -161,4 +161,3 @@ We use:
 - Kacper Stefański
 - Konstantin Panov
 - Piotr Malesa
-

backend/.pre-commit-config.yaml

@@ -1,5 +1,5 @@
 default_language_version:
-  python: python3.11.6
+  python: python3.11.9
 repos:
   - repo: local
     hooks:
@@ -15,14 +15,17 @@ repos:
         entry: isort
         args: ["--check", "--diff"]
         types: [python]
+  - repo: https://github.com/charliermarsh/ruff-pre-commit
+    rev: v0.9.1  
+    hooks:
       - id: ruff
         name: ruff
-        language: system
-        entry: ruff
-        args: ["--force-exclude"]
         types: [python]
+        args: ["--fix"]
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v1.14.1  
+    hooks:
       - id: mypy
         name: mypy
-        language: system
-        entry: mypy
+        args: ["--no-incremental", "--ignore-missing-imports"]
         types: [python]

backend/core/api/api.py

@@ -1,6 +1,5 @@
 import json
 import logging.config
-from typing import List
 
 from fastapi import Body, FastAPI, File, HTTPException, UploadFile
 from fastapi.exceptions import RequestValidationError
@@ -16,7 +15,12 @@
 
 app = FastAPI()
 
-origins = ["http://localhost:3000", "https://wimu-frontend-ccb0bbc023d3.herokuapp.com"]
+origins = [
+    "http://localhost:3000",
+    "https://wimu-frontend-ccb0bbc023d3.herokuapp.com",
+    "https://miditok-visualizer-production-frontend.up.railway.app",
+]
+
 
 app.add_middleware(
     CORSMiddleware, allow_origins=origins, allow_credentials=True, allow_methods=["*"], allow_headers=["*"]
@@ -32,6 +36,11 @@ async def validation_exception_handler(request, exc):
     )
 
 
+def save_to_file(data, filename):  # debug function
+    with open(filename, "w", encoding="utf-8") as f:
+        f.write(str(data))
+
+
 @app.post("/process")
 async def process(config: ConfigModel = Body(...), file: UploadFile = File(...)) -> JSONResponse:
     try:
@@ -40,15 +49,21 @@ async def process(config: ConfigModel = Body(...), file: UploadFile = File(...))
         midi_bytes: bytes = await file.read()
         tokens, notes = tokenize_midi_file(config, midi_bytes)
         serialized_tokens = json.dumps(tokens, cls=TokSequenceEncoder)
-        serialized_notes = [[note.__dict__ for note in track_notes] for track_notes in notes]
+        note_id = 1
+        serialized_notes = []
+        for track_notes in notes:
+            serialized_track = [{**note.__dict__, "note_id": note_id + i} for i, note in enumerate(track_notes)]
+            serialized_notes.append(serialized_track)
+            note_id += len(track_notes)
+
         metrics: MusicInformationData = retrieve_information_from_midi(midi_bytes)
         return JSONResponse(
             content={
                 "success": True,
                 "data": {
                     "tokens": json.loads(serialized_tokens),
                     "notes": serialized_notes,
-                    "metrics": json.loads(metrics.model_dump_json())
+                    "metrics": json.loads(metrics.model_dump_json()),
                 },
                 "error": None,
             }

backend/core/api/logging_middleware.py

@@ -70,6 +70,7 @@ async def _execute_request(self, call_next: Callable, request: Request, request_
 
         except Exception as e:
             self._logger.exception({"path": request.url.path, "method": request.method, "reason": e})
+            return Response(content="Internal Server Error", status_code=500)
 
 
 class AsyncIteratorWrapper:

backend/core/api/model.py

@@ -8,28 +8,35 @@
 
 class ConfigModel(
     BaseModel
-):  # TODO: beat_res, beat_res_rest, chord_maps, chord_tokens_with_root_note, chord_unknown, time_signature_range
-    tokenizer: Literal["REMI", "REMIPlus", "MIDILike", "TSD", "Structured", "CPWord", "Octuple", "MuMIDI", "MMM"]
+):  # TODO: dynamic beat_res, beat_res_rest, chord_maps, chord_tokens_with_root_note, chord_unknown, time_signature_range
+    tokenizer: Literal[
+        "REMI", "REMIPlus", "MIDILike", "TSD", "Structured", "CPWord", "Octuple", "MuMIDI", "MMM", "PerTok"
+    ]
     pitch_range: Annotated[list[Annotated[int, Field(ge=0, le=127)]], Field(min_length=2, max_length=2)]
-    nb_velocities: Annotated[int, Field(ge=0, le=127)]
+    num_velocities: Annotated[int, Field(ge=0, le=127)]
     special_tokens: list[str]
     use_chords: StrictBool
     use_rests: StrictBool
     use_tempos: StrictBool
     use_time_signatures: StrictBool
     use_sustain_pedals: StrictBool
     use_pitch_bends: StrictBool
-    use_programs: StrictBool
     nb_tempos: NonNegativeInt
     tempo_range: Annotated[list[NonNegativeInt], Field(min_length=2, max_length=2)]
     log_tempos: StrictBool
     delete_equal_successive_tempo_changes: StrictBool
     sustain_pedal_duration: StrictBool
     pitch_bend_range: Annotated[list[int], Field(min_length=3, max_length=3)]
     delete_equal_successive_time_sig_changes: StrictBool
+    use_programs: StrictBool
     programs: Optional[Annotated[list[int], Field(min_length=2, max_length=2)]]
     one_token_stream_for_programs: Optional[StrictBool]
     program_changes: Optional[StrictBool]
+    # added for pertok
+    use_microtiming: StrictBool
+    ticks_per_quarter: Annotated[int, Field(ge=24, le=960)]
+    max_microtiming_shift: Annotated[float, Field(ge=0, le=1)]
+    num_microtiming_bins: Annotated[int, Field(ge=1, le=64)]
 
     @model_validator(mode="before")
     @classmethod
@@ -56,12 +63,6 @@ def check_valid_ranges(cls, values):
         if min_pitch_bend > max_pitch_bend:
             raise ValueError("max_pitch_bend must be greater or to than min_pitch_bend")
 
-        # if values.programs is not None:
-        #     min_program = Optional[values.programs[0]]
-        #     max_program = Optional[values.programs[1]]
-        #     if min_program > max_program:
-        #         raise ValueError('max_program must be greater or equal to min_program')
-
         return values
 
 
@@ -100,10 +101,11 @@ class MetricsData:
     empty_beat_rate: float
     drum_pattern_consistency: float
 
+
 @dataclass
 class Note:
     pitch: int
     name: str
     start: int
     end: int
-    velocity: int
+    velocity: int