databricks-tellr-app 0.4.3.dev3403__tar.gz → 0.4.3.dev3432__tar.gz

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: databricks-tellr-app
-Version: 0.4.3.dev3403
+Version: 0.4.3.dev3432
 Summary: Tellr application package for Databricks Apps
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/databricks_tellr_app.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: databricks-tellr-app
-Version: 0.4.3.dev3403
+Version: 0.4.3.dev3432
 Summary: Tellr application package for Databricks Apps
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/databricks_tellr_app.egg-info/SOURCES.txt

@@ -124,7 +124,6 @@ setup.py
 ./src/services/pptx_prompts_defaults.py
 ./src/services/profile_service.py
 ./src/services/streaming_callback.py
-./src/services/svg_sanitizer.py
 ./src/services/validator.py
 ./src/services/converter_jail/__init__.py
 ./src/services/converter_jail/ast_guard.py
@@ -138,6 +137,7 @@ setup.py
 ./src/services/identity_providers/__init__.py
 ./src/services/identity_providers/account_provider.py
 ./src/services/identity_providers/local_provider.py
+./src/services/identity_providers/scim_filter.py
 ./src/services/identity_providers/workspace_provider.py
 ./src/services/tools/__init__.py
 ./src/services/tools/agent_bricks_tool.py
@@ -287,7 +287,6 @@ src/services/pptx_from_records.py
 src/services/pptx_prompts_defaults.py
 src/services/profile_service.py
 src/services/streaming_callback.py
-src/services/svg_sanitizer.py
 src/services/validator.py
 src/services/converter_jail/__init__.py
 src/services/converter_jail/ast_guard.py
@@ -301,6 +300,7 @@ src/services/evaluation/llm_judge.py
 src/services/identity_providers/__init__.py
 src/services/identity_providers/account_provider.py
 src/services/identity_providers/local_provider.py
+src/services/identity_providers/scim_filter.py
 src/services/identity_providers/workspace_provider.py
 src/services/tools/__init__.py
 src/services/tools/agent_bricks_tool.py

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "databricks-tellr-app"
-version = "0.4.3.dev3403"
+version = "0.4.3.dev3432"
 description = "Tellr application package for Databricks Apps"
 readme = "README.md"
 requires-python = ">=3.10"

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/src/services/design_system_service.py

@@ -74,7 +74,6 @@ from src.database.models.design_system import (
     DesignSystemToken,
 )
 from src.services.design_system_compiler import recompute_compiled_style_content
-from src.services.svg_sanitizer import looks_like_svg, sanitize_svg
 
 logger = logging.getLogger(__name__)
 
@@ -2165,20 +2164,6 @@ def _collect_assets_and_files(
                 mime = sniffed
             else:
                 mime = _guess_mime(rel)
-                if mime == "image/svg+xml":
-                    try:
-                        data = sanitize_svg(data)
-                    except ValueError as exc:
-                        raise DesignSystemImportError(
-                            f"Bundle entry '{rel}' is not a valid SVG: {exc}"
-                        ) from exc
-                elif looks_like_svg(data):
-                    # Extension says raster/font/etc. but the bytes are SVG:
-                    # never trust the name to skip sanitization.
-                    raise DesignSystemImportError(
-                        f"Bundle entry '{rel}' contains SVG content but does not "
-                        "have an .svg extension"
-                    )
             width, height = _image_dimensions(data, mime)
             asset = DesignSystemAsset(
                 kind=kind,

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/src/services/identity_provider.py

@@ -14,6 +14,8 @@ from enum import Enum
 from functools import lru_cache
 from typing import Dict, List, Optional
 
+from src.services.identity_providers.scim_filter import escape_scim_filter_string
+
 logger = logging.getLogger(__name__)
 
 
@@ -33,8 +35,9 @@ def _resolve_cached(email: str) -> str:
         provider = get_identity_provider()
         from src.services.identity_providers.workspace_provider import WorkspaceIdentityProvider
         if isinstance(provider._provider, WorkspaceIdentityProvider):
+            escaped_email = escape_scim_filter_string(email)
             for u in provider._provider._client.users.list(
-                filter=f'userName eq "{email}"',
+                filter=f'userName eq "{escaped_email}"',
                 attributes="id,userName,displayName",
             ):
                 if u.display_name:

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/src/services/identity_providers/account_provider.py

@@ -12,6 +12,8 @@ from typing import List, Optional
 
 import requests
 
+from src.services.identity_providers.scim_filter import escape_scim_filter_string
+
 logger = logging.getLogger(__name__)
 
 
@@ -181,9 +183,10 @@ class AccountIdentityProvider:
             Combined list of matching users and groups
         """
         results = []
+        escaped_query = escape_scim_filter_string(query)
         
         if include_users:
-            user_filter = f'userName co "{query}" or displayName co "{query}"'
+            user_filter = f'userName co "{escaped_query}" or displayName co "{escaped_query}"'
             try:
                 users = self.list_users(filter_query=user_filter, max_results=max_results)
                 results.extend(users)
@@ -191,7 +194,7 @@ class AccountIdentityProvider:
                 logger.warning("Failed to search users in Account API")
         
         if include_groups:
-            group_filter = f'displayName co "{query}"'
+            group_filter = f'displayName co "{escaped_query}"'
             try:
                 groups = self.list_groups(filter_query=group_filter, max_results=max_results)
                 results.extend(groups)

databricks_tellr_app-0.4.3.dev3432/src/services/identity_providers/scim_filter.py

@@ -0,0 +1,6 @@
+"""Helpers for safely constructing SCIM filters."""
+
+
+def escape_scim_filter_string(value: str) -> str:
+    """Escape a value for use inside a SCIM filter string literal."""
+    return value.replace("\\", "\\\\").replace('"', '\\"')

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/src/services/identity_providers/workspace_provider.py

@@ -12,6 +12,8 @@ from typing import List, Optional
 
 from databricks.sdk import WorkspaceClient
 
+from src.services.identity_providers.scim_filter import escape_scim_filter_string
+
 logger = logging.getLogger(__name__)
 
 
@@ -50,8 +52,9 @@ class WorkspaceIdentityProvider:
         try:
             kwargs: dict = {"attributes": "id,userName,displayName"}
             if filter_query:
+                escaped_query = escape_scim_filter_string(filter_query)
                 kwargs["filter"] = (
-                    f'userName co "{filter_query}" or displayName co "{filter_query}"'
+                    f'userName co "{escaped_query}" or displayName co "{escaped_query}"'
                 )
 
             users: List[dict] = []
@@ -90,7 +93,8 @@ class WorkspaceIdentityProvider:
         try:
             kwargs: dict = {"attributes": "id,displayName"}
             if filter_query:
-                kwargs["filter"] = f'displayName co "{filter_query}"'
+                escaped_query = escape_scim_filter_string(filter_query)
+                kwargs["filter"] = f'displayName co "{escaped_query}"'
 
             groups: List[dict] = []
             for g in self._client.groups.list(**kwargs):

{databricks_tellr_app-0.4.3.dev3403 → databricks_tellr_app-0.4.3.dev3432}/src/services/image_service.py

@@ -11,7 +11,6 @@ from sqlalchemy.dialects.postgresql import JSONB
 from sqlalchemy.orm import Query, Session
 
 from src.database.models.image import ImageAsset
-from src.services.svg_sanitizer import looks_like_svg, sanitize_svg
 
 logger = logging.getLogger(__name__)
 
@@ -42,20 +41,6 @@ def upload_image(
     if len(file_content) > MAX_FILE_SIZE:
         raise ValueError(f"File too large: {len(file_content)} bytes (max {MAX_FILE_SIZE})")
 
-    # The client-declared MIME type is untrusted: decide on the CONTENT.
-    # SVG can carry script, so it is always sanitized; SVG bytes declared as
-    # any other type are rejected rather than stored under a raster label.
-    if mime_type == "image/svg+xml":
-        try:
-            file_content = sanitize_svg(file_content)
-        except ValueError as exc:
-            raise ValueError(f"Invalid SVG upload: {exc}") from exc
-    elif looks_like_svg(file_content):
-        raise ValueError(
-            f"File content is SVG but was declared as {mime_type}; "
-            "upload it as image/svg+xml."
-        )
-
     # Check for duplicate original_filename among active images (case-insensitive).
     # Ephemeral images (paste-to-chat) skip this check — they use throwaway names
     # and should never block each other.

databricks_tellr_app-0.4.3.dev3403/src/services/svg_sanitizer.py

@@ -1,414 +0,0 @@
-"""Server-side sanitization for SVG images before they are stored."""
-
-from __future__ import annotations
-
-import re
-from urllib.parse import urlsplit
-
-from lxml import etree  # type: ignore[import-untyped]
-
-SVG_NAMESPACE = "http://www.w3.org/2000/svg"
-
-_ALLOWED_ELEMENTS = frozenset(
-    {
-        "svg",
-        "g",
-        "defs",
-        "symbol",
-        "use",
-        "title",
-        "desc",
-        "switch",
-        "view",
-        "path",
-        "rect",
-        "circle",
-        "ellipse",
-        "line",
-        "polyline",
-        "polygon",
-        "text",
-        "tspan",
-        "textPath",
-        "a",
-        "image",
-        "style",
-        "linearGradient",
-        "radialGradient",
-        "stop",
-        "pattern",
-        "marker",
-        "mask",
-        "clipPath",
-        "filter",
-        "feBlend",
-        "feColorMatrix",
-        "feComponentTransfer",
-        "feComposite",
-        "feConvolveMatrix",
-        "feDiffuseLighting",
-        "feDisplacementMap",
-        "feDistantLight",
-        "feDropShadow",
-        "feFlood",
-        "feFuncA",
-        "feFuncB",
-        "feFuncG",
-        "feFuncR",
-        "feGaussianBlur",
-        "feImage",
-        "feMerge",
-        "feMergeNode",
-        "feMorphology",
-        "feOffset",
-        "fePointLight",
-        "feSpecularLighting",
-        "feSpotLight",
-        "feTile",
-        "feTurbulence",
-        "animate",
-        "animateMotion",
-        "animateTransform",
-        "set",
-        "mpath",
-    }
-)
-
-_ALLOWED_ATTRIBUTES = frozenset(
-    {
-        "id",
-        "class",
-        "style",
-        "transform",
-        "transform-origin",
-        "clip-path",
-        "clip-rule",
-        "mask",
-        "filter",
-        "opacity",
-        "fill",
-        "fill-opacity",
-        "fill-rule",
-        "stroke",
-        "stroke-dasharray",
-        "stroke-dashoffset",
-        "stroke-linecap",
-        "stroke-linejoin",
-        "stroke-miterlimit",
-        "stroke-opacity",
-        "stroke-width",
-        "paint-order",
-        "color",
-        "color-interpolation",
-        "color-interpolation-filters",
-        "display",
-        "visibility",
-        "systemLanguage",
-        "requiredFeatures",
-        "requiredExtensions",
-        "lang",
-        "space",
-        "role",
-        "tabindex",
-        "version",
-        "baseProfile",
-        "width",
-        "height",
-        "viewBox",
-        "preserveAspectRatio",
-        "x",
-        "y",
-        "dx",
-        "dy",
-        "zoomAndPan",
-        "d",
-        "points",
-        "rx",
-        "ry",
-        "cx",
-        "cy",
-        "r",
-        "x1",
-        "x2",
-        "y1",
-        "y2",
-        "rotate",
-        "textLength",
-        "lengthAdjust",
-        "font-family",
-        "font-size",
-        "font-style",
-        "font-variant",
-        "font-weight",
-        "letter-spacing",
-        "word-spacing",
-        "text-decoration",
-        "text-anchor",
-        "dominant-baseline",
-        "alignment-baseline",
-        "baseline-shift",
-        "writing-mode",
-        "gradientUnits",
-        "gradientTransform",
-        "spreadMethod",
-        "href",
-        "offset",
-        "stop-color",
-        "stop-opacity",
-        "patternUnits",
-        "patternContentUnits",
-        "patternTransform",
-        "markerWidth",
-        "markerHeight",
-        "markerUnits",
-        "refX",
-        "refY",
-        "orient",
-        "clipPathUnits",
-        "maskUnits",
-        "maskContentUnits",
-        "filterUnits",
-        "primitiveUnits",
-        "in",
-        "in2",
-        "result",
-        "mode",
-        "values",
-        "type",
-        "tableValues",
-        "slope",
-        "intercept",
-        "amplitude",
-        "exponent",
-        "k1",
-        "k2",
-        "k3",
-        "k4",
-        "operator",
-        "radius",
-        "stdDeviation",
-        "xChannelSelector",
-        "yChannelSelector",
-        "scale",
-        "flood-color",
-        "flood-opacity",
-        "kernelMatrix",
-        "divisor",
-        "bias",
-        "targetX",
-        "targetY",
-        "edgeMode",
-        "kernelUnitLength",
-        "preserveAlpha",
-        "lighting-color",
-        "surfaceScale",
-        "diffuseConstant",
-        "specularConstant",
-        "specularExponent",
-        "pointsAtX",
-        "pointsAtY",
-        "pointsAtZ",
-        "limitingConeAngle",
-        "azimuth",
-        "elevation",
-        "seed",
-        "numOctaves",
-        "baseFrequency",
-        "stitchTiles",
-        "attributeName",
-        "attributeType",
-        "begin",
-        "end",
-        "dur",
-        "repeatCount",
-        "repeatDur",
-        "from",
-        "to",
-        "by",
-        "keyTimes",
-        "keySplines",
-        "keyPoints",
-        "calcMode",
-        "additive",
-        "accumulate",
-        "path",
-        "startOffset",
-        "method",
-        "spacing",
-        "crossorigin",
-        "decoding",
-        "loading",
-        "target",
-        "download",
-        "rel",
-    }
-)
-
-_URL_ATTRIBUTES = frozenset({"href"})
-_ANIMATION_ELEMENTS = frozenset({"animate", "animateMotion", "animateTransform", "set"})
-_ANIMATION_FORBIDDEN_TARGETS = frozenset({"href", "style"})
-_SAFE_DATA_IMAGE_TYPES = frozenset({"png", "jpeg", "jpg", "gif", "webp"})
-_UNSAFE_URL_PATTERN = re.compile(
-    r"(?:javascript|vbscript|livescript|mocha):|data:image/svg\+xml",
-    re.IGNORECASE,
-)
-
-
-def _local_name(value: str) -> str:
-    return value.rsplit("}", 1)[-1].rsplit(":", 1)[-1]
-
-
-def _element_local_name(element: etree._Element) -> str:
-    return etree.QName(element).localname
-
-
-def _is_allowed_element(element: etree._Element) -> bool:
-    return _element_local_name(element) in _ALLOWED_ELEMENTS
-
-
-def _normalized_url(value: str) -> str:
-    characters = (character for character in value if character not in "\t\n\r\f\v")
-    return "".join(characters).strip()
-
-
-def _is_safe_url(value: str) -> bool:
-    normalized = _normalized_url(value)
-    if not normalized or normalized.startswith("#"):
-        return True
-
-    parsed = urlsplit(normalized)
-    scheme = parsed.scheme.lower()
-    if not scheme or scheme in {"http", "https", "mailto"}:
-        return True
-    if scheme != "data":
-        return False
-
-    media_type = parsed.path.split(";", 1)[0].lower()
-    image_type = media_type.removeprefix("image/")
-    return image_type in _SAFE_DATA_IMAGE_TYPES
-
-
-def _has_external_use_reference(element: etree._Element) -> bool:
-    return any(
-        _local_name(name) == "href" and not _normalized_url(value).startswith("#")
-        for name, value in element.attrib.items()
-    )
-
-
-def _animation_targets_forbidden_attribute(element: etree._Element) -> bool:
-    target = element.get("attributeName")
-    if target is None:
-        return False
-    local_target = _local_name(target)
-    return local_target.startswith("on") or local_target in _ANIMATION_FORBIDDEN_TARGETS
-
-
-def _contains_unsafe_url(value: str) -> bool:
-    return _UNSAFE_URL_PATTERN.search(_normalized_url(value)) is not None
-
-
-def _sanitize_attributes(root: etree._Element) -> None:
-    for element in root.iter():
-        for name, value in list(element.attrib.items()):
-            local_name = _local_name(name)
-            if local_name not in _ALLOWED_ATTRIBUTES:
-                del element.attrib[name]
-                continue
-            if local_name in _URL_ATTRIBUTES and not _is_safe_url(value):
-                del element.attrib[name]
-            elif local_name == "style" and _contains_unsafe_url(value):
-                del element.attrib[name]
-
-        if _element_local_name(element) == "style" and _contains_unsafe_url(
-            element.text or ""
-        ):
-            element.text = None
-
-
-_RASTER_MAGICS = (
-    b"\x89PNG\r\n\x1a\n",
-    b"\xff\xd8\xff",
-    b"GIF87a",
-    b"GIF89a",
-)
-_SNIFF_WINDOW = 4096
-# An SVG document: optional BOM/whitespace, any XML prolog (declaration, comments,
-# PIs, DOCTYPE with optional internal subset), then ``<svg`` as the ROOT element.
-# Anchoring on the root (not a substring) keeps CSS/HTML/text that merely mentions
-# ``<svg`` (e.g. a ``data:image/svg+xml,<svg ...>`` URI) from being treated as SVG.
-_SVG_ROOT_RE = re.compile(
-    r"\A\ufeff?\s*"
-    r"(?:<\?.*?\?>\s*|<!--.*?-->\s*|<!DOCTYPE(?:[^>\[]|\[.*?\])*>\s*)*"
-    r"<(?:[\w.-]+:)?svg[\s/>]",
-    re.IGNORECASE | re.DOTALL,
-)
-
-
-def looks_like_svg(content: bytes) -> bool:
-    """Content sniff: True if the bytes look like an SVG document.
-
-    Used so a client-declared MIME type (e.g. ``image/png``) is never trusted
-    to decide whether sanitization is needed. Content that starts with a known
-    raster magic number is not SVG; otherwise it is SVG only if ``<svg`` is the
-    root element (after any XML prolog) within the first few KB (UTF-8/ASCII or
-    BOM-marked UTF-16).
-    """
-    head = content[:_SNIFF_WINDOW]
-    if head.startswith(_RASTER_MAGICS):
-        return False
-    if head.startswith((b"\xff\xfe", b"\xfe\xff")):
-        text = head.decode("utf-16", errors="ignore")
-    else:
-        text = head.decode("utf-8", errors="ignore")
-    return _SVG_ROOT_RE.match(text) is not None
-
-
-def sanitize_svg(content: bytes) -> bytes:
-    """Return a sanitized SVG, rejecting malformed or non-SVG XML input."""
-    parser = etree.XMLParser(
-        resolve_entities=False,
-        no_network=True,
-        load_dtd=False,
-        huge_tree=False,
-        remove_comments=True,
-        remove_pis=True,
-    )
-
-    try:
-        root = etree.fromstring(content, parser=parser)
-    except etree.XMLSyntaxError as exc:
-        raise ValueError("SVG must be well-formed XML") from exc
-
-    if root is None:
-        raise ValueError("SVG document is empty")
-
-    root_qname = etree.QName(root)
-    if root_qname.localname != "svg" or root_qname.namespace not in {None, SVG_NAMESPACE}:
-        raise ValueError("SVG root element is required")
-
-    if root.getroottree().docinfo.doctype:
-        raise ValueError("SVG DOCTYPE declarations are not allowed")
-
-    elements_to_remove = []
-    for element in root.iter():
-        if element is root:
-            continue
-
-        local_name = _element_local_name(element)
-        if not _is_allowed_element(element):
-            elements_to_remove.append(element)
-        elif local_name == "use" and _has_external_use_reference(element):
-            elements_to_remove.append(element)
-        elif local_name in _ANIMATION_ELEMENTS and _animation_targets_forbidden_attribute(
-            element
-        ):
-            elements_to_remove.append(element)
-
-    for element in elements_to_remove:
-        parent = element.getparent()
-        if parent is not None:
-            parent.remove(element)
-
-    _sanitize_attributes(root)
-
-    return etree.tostring(root, encoding="utf-8", xml_declaration=True)