Make files into a module of its own.

2 files changed, 53 insertions, 0 deletions

diff --git a/uploader/files/__init__.py b/uploader/files/__init__.py
new file mode 100644
index 0000000..4a62409
--- /dev/null
+++ b/uploader/files/__init__.py
@@ -0,0 +1,4 @@
+from .functions import (fullpath,
+                        save_file,
+                        chunked_binary_read,
+                        sha256_digest_over_file)
diff --git a/uploader/files/functions.py b/uploader/files/functions.py
new file mode 100644
index 0000000..d37a53e
--- /dev/null
+++ b/uploader/files/functions.py
@@ -0,0 +1,49 @@
+"""Utilities to deal with uploaded files."""
+import hashlib
+from pathlib import Path
+from typing import Iterator
+from datetime import datetime
+
+from flask import current_app
+
+from werkzeug.utils import secure_filename
+from werkzeug.datastructures import FileStorage
+
+def save_file(fileobj: FileStorage, upload_dir: Path) -> Path:
+    """Save the uploaded file and return the path."""
+    assert bool(fileobj), "Invalid file object!"
+    hashed_name = hashlib.sha512(
+        f"{fileobj.filename}::{datetime.now().isoformat()}".encode("utf8")
+    ).hexdigest()
+    filename = Path(secure_filename(hashed_name)) # type: ignore[arg-type]
+    if not upload_dir.exists():
+        upload_dir.mkdir()
+
+    filepath = Path(upload_dir, filename)
+    fileobj.save(filepath)
+    return filepath
+
+
+def fullpath(filename: str):
+    """Get a file's full path. This makes use of `flask.current_app`."""
+    return Path(current_app.config["UPLOAD_FOLDER"], filename).absolute()
+
+
+def chunked_binary_read(filepath: Path, chunksize: int = 2048) -> Iterator:
+    """Read a file in binary mode in chunks."""
+    with open(filepath, "rb") as inputfile:
+        while True:
+            data = inputfile.read(chunksize)
+            if data != b"":
+                yield data
+                continue
+            break
+
+
+def sha256_digest_over_file(filepath: Path) -> str:
+    """Compute the sha256 digest over a file's contents."""
+    filehash = hashlib.sha256()
+    for chunk in chunked_binary_read(filepath):
+        filehash.update(chunk)
+
+    return filehash.hexdigest()