Prevent unreasonably long MIME types

Moderation interface is going to use this.

0x0-prune.service

@@ -0,0 +1,22 @@
+[Unit]
+Description=Prune 0x0 files
+After=remote-fs.target
+
+[Service]
+Type=oneshot
+User=nullptr
+WorkingDirectory=/path/to/0x0
+BindPaths=/path/to/0x0
+
+Environment=FLASK_APP=fhost
+ExecStart=/usr/bin/flask prune
+ProtectProc=noaccess
+ProtectSystem=strict
+ProtectHome=tmpfs
+PrivateTmp=true
+PrivateUsers=true
+ProtectKernelLogs=true
+LockPersonality=true
+
+[Install]
+WantedBy=multi-user.target

0x0-prune.timer

@@ -0,0 +1,9 @@
+[Unit]
+Description=Prune 0x0 files
+
+[Timer]
+OnCalendar=hourly
+Persistent=true
+
+[Install]
+WantedBy=timers.target

0x0-vscan.service

@@ -0,0 +1,22 @@
+[Unit]
+Description=Scan 0x0 files with ClamAV
+After=remote-fs.target clamd.service
+
+[Service]
+Type=oneshot
+User=nullptr
+WorkingDirectory=/path/to/0x0
+BindPaths=/path/to/0x0
+
+Environment=FLASK_APP=fhost
+ExecStart=/usr/bin/flask vscan
+ProtectProc=noaccess
+ProtectSystem=strict
+ProtectHome=tmpfs
+PrivateTmp=true
+PrivateUsers=true
+ProtectKernelLogs=true
+LockPersonality=true
+
+[Install]
+WantedBy=multi-user.target

0x0-vscan.timer

@@ -0,0 +1,9 @@
+[Unit]
+Description=Scan 0x0 files with ClamAV
+
+[Timer]
+OnCalendar=hourly
+Persistent=true
+
+[Install]
+WantedBy=timers.target

README.rst

@@ -7,13 +7,13 @@ This is a no-bullshit file hosting and URL shortening service that also runs
 Configuration
 -------------
 
-To configure 0x0, create ``instance/config.py``.
+To configure 0x0, copy ``instance/config.example.py`` to ``instance/config.py``, then edit
-The defaults are at the start of ``fhost.py``. To change them,
+it.   Resonable defaults are set, but there's a couple options you'll need to change
-add them to ``instance/config.py``— for example::
+before running 0x0 for the first time.
 
-    SQLALCHEMY_DATABASE_URI = "sqlite:///some/path/db.sqlite"
+By default, the configuration is stored in the Flask instance directory.
-
+Normally, this is in `./instance`, but it might be different for your system.
-For more information on instance configuration, see
+For details, see
 `the Flask documentation <https://flask.palletsprojects.com/en/2.0.x/config/#instance-folders>`_.
 
 To customize the home and error pages, simply create a ``templates`` directory
@@ -35,10 +35,17 @@ downsides, one of them being that range requests will not work. This is a
 problem for example when streaming media files: It won’t be possible to seek,
 and some ISOBMFF (MP4) files will not play at all.
 
-To make files expire, simply create a cronjob that runs ``cleanup.py`` every
+To make files expire, simply run ``FLASK_APP=fhost flask prune`` every
-now and then.
+now and then. You can use the provided systemd unit files for this::
 
-Before running the service for the first time, run ``FLASK_APP=fhost flask db upgrade``.
+    0x0-prune.service
+    0x0-prune.timer
+
+Make sure to edit them to match your system configuration. In particular,
+set the user and paths in ``0x0-prune.service``.
+
+Before running the service for the first time and every time you update it
+from this git repository, run ``FLASK_APP=fhost flask db upgrade``.
 
 
 NSFW Detection
@@ -49,7 +56,24 @@ neural network model. This works for images and video files and requires
 the following:
 
 * Caffe Python module (built for Python 3)
-* ``ffmpegthumbnailer`` executable in ``$PATH``
+* `PyAV <https://github.com/PyAV-Org/PyAV>`_
+
+
+Virus Scanning
+--------------
+
+0x0 can scan its files with ClamAV’s daemon. As this can take a long time
+for larger files, this does not happen immediately but instead every time
+you run the ``vscan`` command. It is recommended to configure a systemd
+timer or cronjob to do this periodically. Examples are included::
+
+    0x0-vscan.service
+    0x0-vscan.timer
+
+Remember to adjust your size limits in clamd.conf, including
+``StreamMaxLength``!
+
+This feature requires the `clamd module <https://pypi.org/project/clamd/>`_.
 
 
 Network Security Considerations

cleanup.py

@@ -1,44 +1,8 @@
 #!/usr/bin/env python3
 
-"""
+print("This script has been replaced!!")
-    Copyright © 2020 Mia Herkt
+print("Instead, please run")
-    Licensed under the EUPL, Version 1.2 or - as soon as approved
+print("")
-    by the European Commission - subsequent versions of the EUPL
+print("   $ FLASK_APP=fhost flask prune")
-    (the "License");
+print("")
-    You may not use this work except in compliance with the License.
+exit(1);
-    You may obtain a copy of the license at:
-
-        https://joinup.ec.europa.eu/software/page/eupl
-
-    Unless required by applicable law or agreed to in writing,
-    software distributed under the License is distributed on an
-    "AS IS" basis, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,
-    either express or implied.
-    See the License for the specific language governing permissions
-    and limitations under the License.
-"""
-
-import os
-import sys
-import time
-import datetime
-from fhost import app
-
-os.chdir(os.path.dirname(sys.argv[0]))
-os.chdir(app.config["FHOST_STORAGE_PATH"])
-
-files = [f for f in os.listdir(".")]
-
-maxs = app.config["MAX_CONTENT_LENGTH"]
-mind = 30
-maxd = 365
-
-for f in files:
-    stat = os.stat(f)
-    systime = time.time()
-    age = datetime.timedelta(seconds=(systime - stat.st_mtime)).days
-
-    maxage = mind + (-maxd + mind) * (stat.st_size / maxs - 1) ** 3
-
-    if age >= maxage:
-        os.remove(f)

fhost.py

@@ -22,13 +22,20 @@
 from flask import Flask, abort, make_response, redirect, request, send_from_directory, url_for, Response, render_template
 from flask_sqlalchemy import SQLAlchemy
 from flask_migrate import Migrate
+from sqlalchemy import and_, or_
 from jinja2.exceptions import *
 from jinja2 import ChoiceLoader, FileSystemLoader
 from hashlib import sha256
 from magic import Magic
 from mimetypes import guess_extension
+import click
+import os
 import sys
+import time
+import datetime
+import typing
 import requests
+import secrets
 from validators import url as url_valid
 from pathlib import Path
 
@@ -42,6 +49,7 @@ app.config.update(
     FHOST_USE_X_ACCEL_REDIRECT = True, # expect nginx by default
     FHOST_STORAGE_PATH = "up",
     FHOST_MAX_EXT_LENGTH = 9,
+    FHOST_SECRET_BYTES = 16,
     FHOST_EXT_OVERRIDE = {
         "audio/flac" : ".flac",
         "image/gif" : ".gif",
@@ -63,6 +71,13 @@ app.config.update(
     FHOST_UPLOAD_BLACKLIST = None,
     NSFW_DETECT = False,
     NSFW_THRESHOLD = 0.608,
+    VSCAN_SOCKET = None,
+    VSCAN_QUARANTINE_PATH = "quarantine",
+    VSCAN_IGNORE = [
+        "Eicar-Test-Signature",
+        "PUA.Win.Packer.XmMusicFile",
+    ],
+    VSCAN_INTERVAL = datetime.timedelta(days=7),
     URL_ALPHABET = "DEQhd2uFteibPwq0SWBInTpA_jcZL5GKz3YCR14Ulk87Jors9vNHgfaOmMXy6Vx-",
 )
 
@@ -121,12 +136,23 @@ class File(db.Model):
     addr = db.Column(db.UnicodeText)
     removed = db.Column(db.Boolean, default=False)
     nsfw_score = db.Column(db.Float)
+    expiration = db.Column(db.BigInteger)
+    mgmt_token = db.Column(db.String)
+    secret = db.Column(db.String)
+    last_vscan = db.Column(db.DateTime)
+    size = db.Column(db.BigInteger)
 
-    def __init__(self, sha256, ext, mime, addr):
+    def __init__(self, sha256, ext, mime, addr, expiration, mgmt_token):
         self.sha256 = sha256
         self.ext = ext
         self.mime = mime
         self.addr = addr
+        self.expiration = expiration
+        self.mgmt_token = mgmt_token
+
+    @property
+    def is_nsfw(self) -> bool:
+        return self.nsfw_score and self.nsfw_score > app.config["NSFW_THRESHOLD"]
 
     def getname(self):
         return u"{0}{1}".format(su.enbase(self.id), self.ext)
@@ -134,12 +160,58 @@ class File(db.Model):
     def geturl(self):
         n = self.getname()
 
-        if self.nsfw_score and self.nsfw_score > app.config["NSFW_THRESHOLD"]:
+        if self.is_nsfw:
-            return url_for("get", path=n, _external=True, _anchor="nsfw") + "\n"
+            return url_for("get", path=n, secret=self.secret, _external=True, _anchor="nsfw") + "\n"
         else:
-            return url_for("get", path=n, _external=True) + "\n"
+            return url_for("get", path=n, secret=self.secret, _external=True) + "\n"
 
-    def store(file_, addr):
+    def getpath(self) -> Path:
+        return Path(app.config["FHOST_STORAGE_PATH"]) / self.sha256
+
+    def delete(self, permanent=False):
+        self.expiration = None
+        self.mgmt_token = None
+        self.removed = permanent
+        self.getpath().unlink(missing_ok=True)
+
+    # Returns the epoch millisecond that a file should expire
+    #
+    # Uses the expiration time provided by the user (requested_expiration)
+    # upper-bounded by an algorithm that computes the size based on the size of the
+    # file.
+    #
+    # That is, all files are assigned a computed expiration, which can voluntarily
+    # shortened by the user either by providing a timestamp in epoch millis or a
+    # duration in hours.
+    def get_expiration(requested_expiration, size) -> int:
+        current_epoch_millis = time.time() * 1000;
+
+        # Maximum lifetime of the file in milliseconds
+        this_files_max_lifespan = get_max_lifespan(size);
+
+        # The latest allowed expiration date for this file, in epoch millis
+        this_files_max_expiration = this_files_max_lifespan + 1000 * time.time();
+
+        if requested_expiration is None:
+            return this_files_max_expiration
+        elif requested_expiration < 1650460320000:
+            # Treat the requested expiration time as a duration in hours
+            requested_expiration_ms = requested_expiration * 60 * 60 * 1000
+            return min(this_files_max_expiration, current_epoch_millis + requested_expiration_ms)
+        else:
+            # Treat the requested expiration time as a timestamp in epoch millis
+            return min(this_files_max_expiration, requested_expiration)
+
+    """
+    requested_expiration can be:
+        - None, to use the longest allowed file lifespan
+        - a duration (in hours) that the file should live for
+        - a timestamp in epoch millis that the file should expire at
+
+    Any value greater that the longest allowed file lifespan will be rounded down to that
+    value.
+    """
+    def store(file_, requested_expiration: typing.Optional[int], addr, secret: bool):
         data = file_.read()
         digest = sha256(data).hexdigest()
 
@@ -155,6 +227,9 @@ class File(db.Model):
             if mime in app.config["FHOST_MIME_BLACKLIST"] or guess in app.config["FHOST_MIME_BLACKLIST"]:
                 abort(415)
 
+            if len(mime) > 128:
+                abort(400)
+
             if mime.startswith("text/") and not "charset" in mime:
                 mime += "; charset=utf-8"
 
@@ -162,6 +237,8 @@ class File(db.Model):
 
         def get_ext(mime):
             ext = "".join(Path(file_.filename).suffixes[-2:])
+            if len(ext) > app.config["FHOST_MAX_EXT_LENGTH"]:
+                ext = Path(file_.filename).suffixes[-1]
             gmime = mime.split(";")[0]
             guess = guess_extension(gmime)
 
@@ -175,18 +252,38 @@ class File(db.Model):
 
             return ext[:app.config["FHOST_MAX_EXT_LENGTH"]] or ".bin"
 
-        f = File.query.filter_by(sha256=digest).first()
+        expiration = File.get_expiration(requested_expiration, len(data))
+        isnew = True
 
+        f = File.query.filter_by(sha256=digest).first()
         if f:
+            # If the file already exists
             if f.removed:
+                # The file was removed by moderation, so don't accept it back
                 abort(451)
+            if f.expiration is None:
+                # The file has expired, so give it a new expiration date
+                f.expiration = expiration
+
+                # Also generate a new management token
+                f.mgmt_token = secrets.token_urlsafe()
+            else:
+                # The file already exists, update the expiration if needed
+                f.expiration = max(f.expiration, expiration)
+                isnew = False
         else:
             mime = get_mime()
             ext = get_ext(mime)
-            f = File(digest, ext, mime, addr)
+            mgmt_token = secrets.token_urlsafe()
+            f = File(digest, ext, mime, addr, expiration, mgmt_token)
 
         f.addr = addr
 
+        if isnew:
+            f.secret = None
+            if secret:
+                f.secret = secrets.token_urlsafe(app.config["FHOST_SECRET_BYTES"])
+
         storage = Path(app.config["FHOST_STORAGE_PATH"])
         storage.mkdir(parents=True, exist_ok=True)
         p = storage / digest
@@ -194,16 +291,15 @@ class File(db.Model):
         if not p.is_file():
             with open(p, "wb") as of:
                 of.write(data)
-        else:
+
-            p.touch()
+        f.size = len(data)
 
         if not f.nsfw_score and app.config["NSFW_DETECT"]:
             f.nsfw_score = nsfw.detect(p)
 
         db.session.add(f)
         db.session.commit()
-        return f
+        return f, isnew
-
 
 
 class UrlEncoder(object):
@@ -260,15 +356,30 @@ def in_upload_bl(addr):
 
     return False
 
-def store_file(f, addr):
+"""
+requested_expiration can be:
+    - None, to use the longest allowed file lifespan
+    - a duration (in hours) that the file should live for
+    - a timestamp in epoch millis that the file should expire at
+
+Any value greater that the longest allowed file lifespan will be rounded down to that
+value.
+"""
+def store_file(f, requested_expiration:  typing.Optional[int], addr, secret: bool):
     if in_upload_bl(addr):
         return "Your host is blocked from uploading files.\n", 451
 
-    sf = File.store(f, addr)
+    sf, isnew = File.store(f, requested_expiration, addr, secret)
 
-    return sf.geturl()
+    response = make_response(sf.geturl())
+    response.headers["X-Expires"] = sf.expiration
 
-def store_url(url, addr):
+    if isnew:
+        response.headers["X-Token"] = sf.mgmt_token
+
+    return response
+
+def store_url(url, addr, secret: bool):
     if is_fhost_url(url):
         abort(400)
 
@@ -289,40 +400,81 @@ def store_url(url, addr):
 
             f = urlfile(read=r.raw.read, content_type=r.headers["content-type"], filename="")
 
-            return store_file(f, addr)
+            return store_file(f, None, addr, secret)
         else:
             abort(413)
     else:
         abort(411)
 
-@app.route("/<path:path>")
+def manage_file(f):
-def get(path):
+    try:
-    path = Path(path.split("/", 1)[0])
+        assert(request.form["token"] == f.mgmt_token)
-    sufs = "".join(path.suffixes[-2:])
+    except:
-    name = path.name[:-len(sufs) or None]
+        abort(401)
+
+    if "delete" in request.form:
+        f.delete()
+        db.session.commit()
+        return ""
+    if "expires" in request.form:
+        try:
+            requested_expiration = int(request.form["expires"])
+        except ValueError:
+            abort(400)
+
+        f.expiration = File.get_expiration(requested_expiration, f.size)
+        db.session.commit()
+        return "", 202
+
+    abort(400)
+
+@app.route("/<path:path>", methods=["GET", "POST"])
+@app.route("/s/<secret>/<path:path>", methods=["GET", "POST"])
+def get(path, secret=None):
+    p = Path(path.split("/", 1)[0])
+    sufs = "".join(p.suffixes[-2:])
+    name = p.name[:-len(sufs) or None]
+
+    if "." in name:
+        abort(404)
+
     id = su.debase(name)
 
     if sufs:
         f = File.query.get(id)
 
         if f and f.ext == sufs:
+            if f.secret != secret:
+                abort(404)
+
             if f.removed:
                 abort(451)
 
-            fpath = Path(app.config["FHOST_STORAGE_PATH"]) / f.sha256
+            fpath = f.getpath()
 
             if not fpath.is_file():
                 abort(404)
 
+            if request.method == "POST":
+                return manage_file(f)
+
             if app.config["FHOST_USE_X_ACCEL_REDIRECT"]:
                 response = make_response()
                 response.headers["Content-Type"] = f.mime
-                response.headers["Content-Length"] = fpath.stat().st_size
+                response.headers["Content-Length"] = f.size
                 response.headers["X-Accel-Redirect"] = "/" + str(fpath)
+            else:
+                response = send_from_directory(app.config["FHOST_STORAGE_PATH"], f.sha256, mimetype = f.mime)
+
+            response.headers["X-Expires"] = f.expiration
             return response
     else:
-                return send_from_directory(app.config["FHOST_STORAGE_PATH"], f.sha256, mimetype = f.mime)
+        if request.method == "POST":
-    else:
+            abort(405)
+
+        if "/" in path:
+            abort(404)
+
         u = URL.query.get(id)
 
         if u:
@@ -334,11 +486,34 @@ def get(path):
 def fhost():
     if request.method == "POST":
         sf = None
+        secret = "secret" in request.form
 
         if "file" in request.files:
-            return store_file(request.files["file"], request.remote_addr)
+            try:
+                # Store the file with the requested expiration date
+                return store_file(
+                    request.files["file"],
+                    int(request.form["expires"]),
+                    request.remote_addr,
+                    secret
+                )
+            except ValueError:
+                # The requested expiration date wasn't properly formed
+                abort(400)
+            except KeyError:
+                # No expiration date was requested, store with the max lifespan
+                return store_file(
+                    request.files["file"],
+                    None,
+                    request.remote_addr,
+                    secret
+                )
         elif "url" in request.form:
-            return store_url(request.form["url"], request.remote_addr)
+            return store_url(
+                request.form["url"],
+                request.remote_addr,
+                secret
+            )
         elif "shorten" in request.form:
             return shorten(request.form["shorten"])
 
@@ -353,6 +528,7 @@ Disallow: /
 """
 
 @app.errorhandler(400)
+@app.errorhandler(401)
 @app.errorhandler(404)
 @app.errorhandler(411)
 @app.errorhandler(413)
@@ -361,6 +537,129 @@ Disallow: /
 @app.errorhandler(451)
 def ehandler(e):
     try:
-        return render_template(f"{e.code}.html", id=id), e.code
+        return render_template(f"{e.code}.html", id=id, request=request), e.code
     except TemplateNotFound:
         return "Segmentation fault\n", e.code
+
+@app.cli.command("prune")
+def prune():
+    """
+    Clean up expired files
+
+    Deletes any files from the filesystem which have hit their expiration time.  This
+    doesn't remove them from the database, only from the filesystem.  It's recommended
+    that server owners run this command regularly, or set it up on a timer.
+    """
+    current_time = time.time() * 1000;
+
+    # The path to where uploaded files are stored
+    storage = Path(app.config["FHOST_STORAGE_PATH"])
+
+    # A list of all files who've passed their expiration times
+    expired_files = File.query\
+        .where(
+            and_(
+                File.expiration.is_not(None),
+                File.expiration < current_time
+            )
+        )
+
+    files_removed = 0;
+
+    # For every expired file...
+    for file in expired_files:
+        # Log the file we're about to remove
+        file_name = file.getname()
+        file_hash = file.sha256
+        file_path = storage / file_hash
+        print(f"Removing expired file {file_name} [{file_hash}]")
+
+        # Remove it from the file system
+        try:
+            os.remove(file_path)
+            files_removed += 1;
+        except FileNotFoundError:
+            pass # If the file was already gone, we're good
+        except OSError as e:
+            print(e)
+            print(
+                "\n------------------------------------"
+                "Encountered an error while trying to remove file {file_path}.  Double"
+                "check to make sure the server is configured correctly, permissions are"
+                "okay, and everything is ship shape, then try again.")
+            return;
+
+        # Finally, mark that the file was removed
+        file.expiration = None;
+    db.session.commit()
+
+    print(f"\nDone!  {files_removed} file(s) removed")
+
+""" For a file of a given size, determine the largest allowed lifespan of that file
+
+Based on the current app's configuration:  Specifically, the MAX_CONTENT_LENGTH, as well
+as FHOST_{MIN,MAX}_EXPIRATION.
+
+This lifespan may be shortened by a user's request, but no files should be allowed to
+expire at a point after this number.
+
+Value returned is a duration in milliseconds.
+"""
+def get_max_lifespan(filesize: int) -> int:
+    min_exp = app.config.get("FHOST_MIN_EXPIRATION", 30 * 24 * 60 * 60 * 1000)
+    max_exp = app.config.get("FHOST_MAX_EXPIRATION", 365 * 24 * 60 * 60 * 1000)
+    max_size = app.config.get("MAX_CONTENT_LENGTH", 256 * 1024 * 1024)
+    return min_exp + int((-max_exp + min_exp) * (filesize / max_size - 1) ** 3)
+
+def do_vscan(f):
+    if f["path"].is_file():
+        with open(f["path"], "rb") as scanf:
+            try:
+                f["result"] = list(app.config["VSCAN_SOCKET"].instream(scanf).values())[0]
+            except:
+                f["result"] = ("SCAN FAILED", None)
+    else:
+        f["result"] = ("FILE NOT FOUND", None)
+
+    return f
+
+@app.cli.command("vscan")
+def vscan():
+    if not app.config["VSCAN_SOCKET"]:
+        print("""Error: Virus scanning enabled but no connection method specified.
+Please set VSCAN_SOCKET.""")
+        sys.exit(1)
+
+    qp = Path(app.config["VSCAN_QUARANTINE_PATH"])
+    qp.mkdir(parents=True, exist_ok=True)
+
+    from multiprocessing import Pool
+    with Pool() as p:
+        if isinstance(app.config["VSCAN_INTERVAL"], datetime.timedelta):
+            scandate = datetime.datetime.now() - app.config["VSCAN_INTERVAL"]
+            res = File.query.filter(or_(File.last_vscan < scandate,
+                                        File.last_vscan == None),
+                                    File.removed == False)
+        else:
+            res = File.query.filter(File.last_vscan == None, File.removed == False)
+
+        work = [{"path" : f.getpath(), "name" : f.getname(), "id" : f.id} for f in res]
+
+        results = []
+        for i, r in enumerate(p.imap_unordered(do_vscan, work)):
+            if r["result"][0] != "OK":
+                print(f"{r['name']}: {r['result'][0]} {r['result'][1] or ''}")
+
+            found = False
+            if r["result"][0] == "FOUND":
+                if not r["result"][1] in app.config["VSCAN_IGNORE"]:
+                    r["path"].rename(qp / r["name"])
+                    found = True
+
+            results.append({
+                "id" : r["id"],
+                "last_vscan" : None if r["result"][0] == "SCAN FAILED" else datetime.datetime.now(),
+                "removed" : found})
+
+        db.session.bulk_update_mappings(File, results)
+        db.session.commit()

instance/config.example.py

@@ -0,0 +1,215 @@
+
+
+ ################################################################################
+ # This is a configuration file for 0x0 / The Null Pointer                      #
+ #                                                                              #
+ # The default values here are set to generally reasonable defaults, but a      #
+ # couple of things need your attention.  Specifically, make sure you set       #
+ # SQLALCHEMY_DATABASE_URI.  You'll also probably want to configure             #
+ # FHOST_USE_X_SENDFILE and FHOST_USE_X_ACCEL_REDIRECT to match your webserver. #
+ #                                                                              #
+ # Need help, or find anything confusing?  Try opening up an issue!             #
+ # https://git.0x0.st/mia/0x0/issues/new                                        #
+ ################################################################################
+
+
+
+# The database URL for the database 0x0 should use
+#
+# See https://docs.sqlalchemy.org/en/20/core/engines.html#backend-specific-urls
+# for help configuring these for your database.
+#
+# For small and medium servers, it's plenty sufficient to just use an sqlite
+# database.  In this case, the database URI you want to use is just
+#
+# sqlite:/// + /path/to/your/database.db
+#
+# Until https://git.0x0.st/mia/0x0/issues/70 is resolved, it's recommended that
+# any sqlite databases use an absolute path, as relative paths aren't consistently
+# resolved.
+SQLALCHEMY_DATABASE_URI = 'sqlite:///' + '/path/to/database.sqlite'
+
+
+# The maximum allowable upload size, in bytes
+#
+# Keep in mind that this affects the expiration of files as well!  The closer a
+# file is to the max content length, the less time it will last before being
+# deleted.
+MAX_CONTENT_LENGTH = 256 * 1024 * 1024 # Default: 256MiB
+
+
+# The maximum length of URLs we'll shorten, in characters
+#
+# If a user tries to submit a URL longer than this, we'll reject their request
+# with a 414 REQUEST URI TOO LONG.
+MAX_URL_LENGTH = 4096
+
+
+# The minimum and maximum amount of time we'll retain a file for
+#
+# Small files (nearing zero bytes) are stored for the longest possible expiration date,
+# while larger files (nearing MAX_CONTENT_LENGTH bytes) are stored for the shortest amount
+# of time.  Values between these two extremes are interpolated with an exponential curve,
+# like the one shown on the index page.
+#
+# All times are in milliseconds.  If you want all files to be stored for the same amount
+# of time, set these to the same value.
+FHOST_MIN_EXPIRATION = 30  * 24 * 60 * 60 * 1000
+FHOST_MAX_EXPIRATION = 365 * 24 * 60 * 60 * 1000
+
+
+# Use the X-SENDFILE header to speed up serving files w/ compatible webservers
+#
+# Some webservers can be configured use the X-Sendfile header to handle sending
+# large files on behalf of the application.  If your server is setup to do
+# this, set this variable to True
+USE_X_SENDFILE = False
+
+
+# Use X-Accel-Redirect to speed up serving files w/ compatible webservers
+#
+# Other webservers, like nginx and Caddy, use the X-Accel-Redirect header to
+# accomplish a very similar thing to X-Sendfile (above).  If your webserver is
+# configured to do this, set this variable to True
+#
+# Note:  It's recommended that you use either X-Sendfile or X-Accel-Redirect
+# when you deploy in production.
+FHOST_USE_X_ACCEL_REDIRECT = True # expect nginx by default
+
+
+# The directory that 0x0 should store uploaded files in
+#
+# Whenever a file is uploaded to 0x0, we store it here!  Relative paths are
+# resolved relative to the working directory that 0x0 is being run from.
+FHOST_STORAGE_PATH = "up"
+
+
+# The maximum acceptable user-specified file extension
+#
+# When a user uploads a file, in most cases, we keep the file extension they
+# provide.  But!  If the specified file extension is longer than
+# FHOST_MAX_EXT_LENGTH, we truncate it.  So if a user tries to upload the file
+# "myfile.withareallongext", but FHOST_MAX_EXT_LENGTH is set to 9, then the
+# extension that we keep is ".withareal"
+FHOST_MAX_EXT_LENGTH = 9
+
+
+# The number of bytes used for "secret" URLs
+#
+# When a user uploads a file with the "secret" option, 0x0 generates a string
+# from this many bytes of random data. It is base64-encoded, so on average
+# each byte results in approximately 1.3 characters.
+FHOST_SECRET_BYTES = 16
+
+# A list of filetypes to use when the uploader doesn't specify one
+#
+# When a user uploads a file with no file extension, we try to find an extension that
+# works for that file.  This configuration option is the first thing that we check.  If
+# the type of a file without an extension is in this dict, then it'll be used as the file
+# extension for that file. Otherwise, we try to pick something sensible from libmagic's
+# database.
+#
+# For example, if the user uploads "myfile" with no extension, and the file is a jpeg
+# image, the file will get a URL like "eAa.jpg"
+#
+# For a list of MIME types you can use in this list, check
+# https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_types/Common_types
+FHOST_EXT_OVERRIDE = {
+    "audio/flac" : ".flac",
+    "image/gif" : ".gif",
+    "image/jpeg" : ".jpg",
+    "image/png" : ".png",
+    "image/svg+xml" : ".svg",
+    "video/webm" : ".webm",
+    "video/x-matroska" : ".mkv",
+    "application/octet-stream" : ".bin",
+    "text/plain" : ".log",
+    "text/plain" : ".txt",
+    "text/x-diff" : ".diff",
+}
+
+
+# Control which files aren't allowed to be uploaded
+#
+# Certain kinds of files are never accepted.  If the file claims to be one of
+# these types of files, or if we look at the contents of the file and it looks
+# like one of these filetypes, then we reject the file outright with a 415
+# UNSUPPORTED MEDIA EXCEPTION
+FHOST_MIME_BLACKLIST = [
+    "application/x-dosexec",
+    "application/java-archive",
+    "application/java-vm"
+]
+
+
+# A list of IP addresses which are blacklisted from uploading files
+#
+# Can be set to the path of a file with an IP address on each line.  The file
+# can also include comment lines using a pound sign (#).  Paths are resolved
+# relative to the instance/ directory.
+#
+# If this is set to None, then no IP blacklist will be consulted.
+FHOST_UPLOAD_BLACKLIST = None
+
+
+# Enables support for detecting NSFW images
+#
+# Consult README.md for additional dependencies before setting to True
+NSFW_DETECT = False
+
+
+# The cutoff for when an image is considered NFSW
+#
+# When the NSFW detection algorithm generates an output higher than this
+# number, an image is considered to be NSFW.  NSFW images aren't declined, but
+# are marked as NSFW.
+#
+# If NSFW_DETECT is set to False, then this has no effect.
+NSFW_THRESHOLD = 0.608
+
+
+# If you want to scan files for viruses using ClamAV, specify the socket used
+# for connections here. You will need the clamd module.
+# Since this can take a very long time on larger files, it is not done
+# immediately but every time you run the vscan command. It is recommended to
+# configure a systemd timer or cronjob to do this periodically.
+# Remember to adjust your size limits in clamd.conf, including StreamMaxLength!
+#
+# Example:
+# from clamd import ClamdUnixSocket
+# VSCAN_SOCKET = ClamdUnixSocket("/run/clamav/clamd-socket")
+
+# This is the directory that files flagged as malicious are moved to.
+# Relative paths are resolved relative to the working directory
+# of the 0x0 process.
+VSCAN_QUARANTINE_PATH = "quarantine"
+
+# Since updated virus definitions might catch some files that were previously
+# reported as clean, you may want to rescan old files periodically.
+# Set this to a datetime.timedelta to specify the frequency, or None to
+# disable rescanning.
+from datetime import timedelta
+VSCAN_INTERVAL = timedelta(days=7)
+
+# Some files flagged by ClamAV are usually not malicious, especially if the
+# DetectPUA option is enabled in clamd.conf. This is a list of signatures
+# that will be ignored.
+VSCAN_IGNORE = [
+    "Eicar-Test-Signature",
+    "PUA.Win.Packer.XmMusicFile",
+]
+
+# A list of all characters which can appear in a URL
+#
+# If this list is too short, then URLs can very quickly become long.
+# Generally, the default value for this should work for basically all usecases.
+URL_ALPHABET = "DEQhd2uFteibPwq0SWBInTpA_jcZL5GKz3YCR14Ulk87Jors9vNHgfaOmMXy6Vx-"
+
+
+ #################################################################################
+ # CONGRATULATIONS!  You made it all the way through!                            #
+ # If you want to go even further to customize your instance, try checking out   #
+ # the templates in the templates/ directory to customize your landing page, 404 #
+ # page, and other error pages.                                                  #
+ #################################################################################
+

migrations/versions/0659d7b9eea8_.py

@@ -0,0 +1,26 @@
+"""add file management token
+
+Revision ID: 0659d7b9eea8
+Revises: 939a08e1d6e5
+Create Date: 2022-11-30 01:06:53.362973
+
+"""
+
+# revision identifiers, used by Alembic.
+revision = '0659d7b9eea8'
+down_revision = '939a08e1d6e5'
+
+from alembic import op
+import sqlalchemy as sa
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('file', sa.Column('mgmt_token', sa.String(), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('file', 'mgmt_token')
+    # ### end Alembic commands ###

migrations/versions/30bfe33aa328_add_file_size_field.py

@@ -0,0 +1,46 @@
+"""add file size field
+
+Revision ID: 30bfe33aa328
+Revises: 5cee97aab219
+Create Date: 2022-12-13 22:32:12.242394
+
+"""
+
+# revision identifiers, used by Alembic.
+revision = '30bfe33aa328'
+down_revision = '5cee97aab219'
+
+from alembic import op
+import sqlalchemy as sa
+from sqlalchemy.ext.automap import automap_base
+from sqlalchemy.orm import Session
+from flask import current_app
+from pathlib import Path
+
+Base = automap_base()
+
+def upgrade():
+    op.add_column('file', sa.Column('size', sa.BigInteger(), nullable=True))
+    bind = op.get_bind()
+    Base.prepare(autoload_with=bind)
+    File = Base.classes.file
+    session = Session(bind=bind)
+
+    storage = Path(current_app.config["FHOST_STORAGE_PATH"])
+
+    updates = []
+    files = session.scalars(sa.select(File).where(sa.not_(File.removed)))
+    for f in files:
+        p = storage / f.sha256
+        if p.is_file():
+            updates.append({
+                "id" : f.id,
+                "size" : p.stat().st_size
+            })
+
+    session.bulk_update_mappings(File, updates)
+    session.commit()
+
+
+def downgrade():
+    op.drop_column('file', 'size')

migrations/versions/5cee97aab219_.py

@@ -0,0 +1,26 @@
+"""add date of last virus scan
+
+Revision ID: 5cee97aab219
+Revises: e2e816056589
+Create Date: 2022-12-10 16:39:56.388259
+
+"""
+
+# revision identifiers, used by Alembic.
+revision = '5cee97aab219'
+down_revision = 'e2e816056589'
+
+from alembic import op
+import sqlalchemy as sa
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('file', sa.Column('last_vscan', sa.DateTime(), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('file', 'last_vscan')
+    # ### end Alembic commands ###

migrations/versions/939a08e1d6e5_.py

@@ -0,0 +1,81 @@
+"""add file expirations
+
+Revision ID: 939a08e1d6e5
+Revises: 7e246705da6a
+Create Date: 2022-11-22 12:16:32.517184
+
+"""
+
+# revision identifiers, used by Alembic.
+revision = '939a08e1d6e5'
+down_revision = '7e246705da6a'
+
+from alembic import op
+from flask import current_app
+from flask_sqlalchemy import SQLAlchemy
+from pathlib import Path
+import sqlalchemy as sa
+from sqlalchemy.ext.automap import automap_base
+from sqlalchemy.orm import Session
+
+import os
+import time
+
+""" For a file of a given size, determine the largest allowed lifespan of that file
+
+Based on the current app's configuration:  Specifically, the MAX_CONTENT_LENGTH, as well
+as FHOST_{MIN,MAX}_EXPIRATION.
+
+This lifespan may be shortened by a user's request, but no files should be allowed to
+expire at a point after this number.
+
+Value returned is a duration in milliseconds.
+"""
+def get_max_lifespan(filesize: int) -> int:
+    min_exp = current_app.config.get("FHOST_MIN_EXPIRATION", 30 * 24 * 60 * 60 * 1000)
+    max_exp = current_app.config.get("FHOST_MAX_EXPIRATION", 365 * 24 * 60 * 60 * 1000)
+    max_size = current_app.config.get("MAX_CONTENT_LENGTH", 256 * 1024 * 1024)
+    return min_exp + int((-max_exp + min_exp) * (filesize / max_size - 1) ** 3)
+
+Base = automap_base()
+
+def upgrade():
+    op.add_column('file', sa.Column('expiration', sa.BigInteger()))
+
+    bind = op.get_bind()
+    Base.prepare(autoload_with=bind)
+    File = Base.classes.file
+    session = Session(bind=bind)
+
+    storage = Path(current_app.config["FHOST_STORAGE_PATH"])
+    current_time = time.time() * 1000;
+
+    # List of file hashes which have not expired yet
+    # This could get really big for some servers
+    try:
+        unexpired_files = os.listdir(storage)
+    except FileNotFoundError:
+        return # There are no currently unexpired files
+
+    # Calculate an expiration date for all existing files
+    files = session.scalars(
+            sa.select(File)
+            .where(
+                sa.not_(File.removed),
+                File.sha256.in_(unexpired_files)
+            )
+        )
+    updates = [] # We coalesce updates to the database here
+    for file in files:
+        file_path = storage / file.sha256
+        stat = os.stat(file_path)
+        max_age = get_max_lifespan(stat.st_size) # How long the file is allowed to live, in ms
+        file_birth = stat.st_mtime * 1000 # When the file was created, in ms
+        updates.append({'id': file.id, 'expiration': int(file_birth + max_age)})
+
+    # Apply coalesced updates
+    session.bulk_update_mappings(File, updates)
+    session.commit()
+
+def downgrade():
+    op.drop_column('file', 'expiration')

migrations/versions/e2e816056589_.py

@@ -0,0 +1,26 @@
+"""add URL secret
+
+Revision ID: e2e816056589
+Revises: 0659d7b9eea8
+Create Date: 2022-12-01 02:16:15.976864
+
+"""
+
+# revision identifiers, used by Alembic.
+revision = 'e2e816056589'
+down_revision = '0659d7b9eea8'
+
+from alembic import op
+import sqlalchemy as sa
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('file', sa.Column('secret', sa.String(), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('file', 'secret')
+    # ### end Alembic commands ###

nsfw_detect.py

@@ -22,21 +22,21 @@ import numpy as np
 import os
 import sys
 from io import BytesIO
-from subprocess import run, PIPE, DEVNULL
+from pathlib import Path
-
-import caffe
 
 os.environ["GLOG_minloglevel"] = "2"  # seriously :|
-
+import caffe
+import av
+av.logging.set_level(av.logging.PANIC)
 
 class NSFWDetector:
     def __init__(self):
-
+        npath = Path(__file__).parent / "nsfw_model"
-        npath = os.path.join(os.path.dirname(__file__), "nsfw_model")
         self.nsfw_net = caffe.Net(
-            os.path.join(npath, "deploy.prototxt"),
+            str(npath / "deploy.prototxt"),
-            os.path.join(npath, "resnet_50_1by2_nsfw.caffemodel"),
+            caffe.TEST,
-            caffe.TEST)
+            weights = str(npath / "resnet_50_1by2_nsfw.caffemodel")
+        )
         self.caffe_transformer = caffe.io.Transformer({
             'data': self.nsfw_net.blobs['data'].data.shape
         })
@@ -50,7 +50,7 @@ class NSFWDetector:
         self.caffe_transformer.set_channel_swap('data', (2, 1, 0))
 
     def _compute(self, img):
-        image = caffe.io.load_image(BytesIO(img))
+        image = caffe.io.load_image(img)
 
         H, W, _ = image.shape
         _, _, h, w = self.nsfw_net.blobs["data"].data.shape
@@ -72,16 +72,26 @@ class NSFWDetector:
 
     def detect(self, fpath):
         try:
-            ff = run([
+            with av.open(fpath) as container:
-                "ffmpegthumbnailer", "-m", "-o-", "-s256", "-t50%", "-a",
+                try: container.seek(int(container.duration / 2))
-                "-cpng", "-i", fpath
+                except: container.seek(0)
-            ], stdout=PIPE, stderr=DEVNULL, check=True)
+
-            image_data = ff.stdout
+                frame = next(container.decode(video=0))
+
+                if frame.width >= frame.height:
+                    w = 256
+                    h = int(frame.height * (256 / frame.width))
+                else:
+                    w = int(frame.width * (256 / frame.height))
+                    h = 256
+                frame = frame.reformat(width=w, height=h, format="rgb24")
+                img = BytesIO()
+                frame.to_image().save(img, format="ppm")
+
+            scores = self._compute(img)
         except:
             return -1.0
 
-        scores = self._compute(image_data)
-
         return scores[1]
 
 

templates/401.html

@@ -0,0 +1,2 @@
+rm: cannot remove '{{ request.path.split("/")[1] }}': Permission denied
+

templates/index.html

@@ -6,11 +6,35 @@ HTTP POST files here:
     curl -F'file=@yourfile.png' {{ fhost_url }}
 You can also POST remote URLs:
     curl -F'url=http://example.com/image.jpg' {{ fhost_url }}
+If you don't want the resulting URL to be easy to guess:
+    curl -F'file=@yourfile.png' -Fsecret= {{ fhost_url }}
+    curl -F'url=http://example.com/image.jpg' -Fsecret= {{ fhost_url }}
 Or you can shorten URLs:
     curl -F'shorten=http://example.com/some/long/url' {{ fhost_url }}
 
+It is possible to append your own file name to the URL:
+    {{ fhost_url }}/aaa.jpg/image.jpeg
+
 File URLs are valid for at least 30 days and up to a year (see below).
 Shortened URLs do not expire.
+
+Files can be set to expire sooner by adding an "expires" parameter (in hours)
+    curl -F'file=@yourfile.png' -Fexpires=24 {{ fhost_url }}
+OR by setting "expires" to a timestamp in epoch milliseconds
+    curl -F'file=@yourfile.png' -Fexpires=1681996320000 {{ fhost_url }}
+
+Expired files won't be removed immediately, but will be removed as part of
+the next purge.
+
+Whenever a file that does not already exist or has expired is uploaded,
+the HTTP response header includes an X-Token field. You can use this
+to perform management operations on the file.
+
+To delete the file immediately:
+    curl -Ftoken=token_here -Fdelete= {{ fhost_url }}/abc.txt
+To change the expiration date (see above):
+    curl -Ftoken=token_here -Fexpires=3 {{ fhost_url }}/abc.txt
+
 {% set max_size = config["MAX_CONTENT_LENGTH"]|filesizeformat(True) %}
 Maximum file size: {{ max_size }}
 Not allowed: {{ config["FHOST_MIME_BLACKLIST"]|join(", ") }}
@@ -22,24 +46,24 @@ FILE RETENTION PERIOD
 retention = min_age + (-max_age + min_age) * pow((file_size / max_size - 1), 3)
 
    days
-    365 |  \\
+ {{'{: 6}'.format(config.get("FHOST_MAX_EXPIRATION", 31536000000)//86400000)}} |  \
-        |   \\
+        |   \
-        |    \\
+        |    \
-        |     \\
+        |     \
-        |      \\
+        |      \
-        |       \\
+        |       \
         |        ..
-        |          \\
+        |          \
-  197.5 | ----------..-------------------------------------------
+ {{'{: 6.1f}'.format((config.get("FHOST_MIN_EXPIRATION", 2592000000)/2 + config.get("FHOST_MAX_EXPIRATION", 31536000000)/2)/86400000)}} | ----------..-------------------------------------------
         |             ..
-        |               \\
+        |               \
         |                ..
         |                  ...
         |                     ..
         |                       ...
         |                          ....
         |                              ......
-     30 |                                    ....................
+ {{'{: 6}'.format(config.get("FHOST_MIN_EXPIRATION", 2592000000)//86400000)}} |                                    ....................
           0{{ ((config["MAX_CONTENT_LENGTH"]/2)|filesizeformat(True)).split(" ")[0].rjust(27) }}{{ max_size.split(" ")[0].rjust(27) }}
            {{ max_size.split(" ")[1].rjust(54) }}
 </pre>

tests/test_client.py

@@ -56,8 +56,6 @@ def test_client(client):
         ]),
         (302, [
             "E",
-            "E/test",
-            "E/test.bin",
         ]),
         (404, [
             "test.bin",
@@ -67,6 +65,8 @@ def test_client(client):
             "test/test",
             "test.bin/test.py",
             "E.bin",
+            "E/test",
+            "E/test.bin",
         ]),
         (451, [
             "Q.truncate",