added meta for all files, html preview

akarpov/files/previews/__init__.py

@@ -34,7 +34,7 @@
     },
     "text": {
         "css": text.common.view,
-        "html": text.common.view,
+        "html": text.html.view,
         "javascript": text.common.view,
         "plain": text.plain.view,
         "csv": text.csv.view,
@@ -76,6 +76,7 @@
         "mpeg": video.mp4.view,
         "oga": audio.oga.view,
         "pdf": application.pdf.view,
+        "html": text.html.view,
     }
     | source_code
     | fonts_ext
@@ -88,12 +89,23 @@
         "ogg": video.basic.meta,
         "mpeg": video.basic.meta,
         "quicktime": video.basic.meta,
-    }
+    },
+    "text": {
+        "css": text.common.meta,
+        "html": text.common.meta,
+        "javascript": text.common.meta,
+        "plain": text.common.meta,
+    },
 }
 
+
+meta_source_code = {}
+for ext in text.common.language_previews.keys():
+    source_code[ext] = text.common.meta
+
 meta_extensions = {
     "mp4": video.basic.meta,
     "mov": video.basic.meta,
     "ogv": video.basic.meta,
     "mpeg": video.basic.meta,
-}
+} | meta_source_code

akarpov/files/previews/text/__init__.py

@@ -1 +1 @@
-from . import common, csv, plain  # noqa
+from . import common, csv, html, plain  # noqa

akarpov/files/previews/text/common.py

@@ -26,7 +26,6 @@
     "pl": "perl",
     "PL": "perl",
     "htm": "html",
-    "html": "html",
 }
 
 
@@ -71,20 +70,27 @@ def meta(file: File):
     with file.file.open("r") as f:
         lines = f.readlines()
     for line in lines:
-        descr += line + "\n"
+        if i == 0:
+            descr += line + "\n"
+        else:
+            descr += line + "    "
         i += 1
-        if i > 100:
+        if i > 20:
             descr += "..."
             break
     url = file.get_absolute_url()
+    section = ""
+    if file.file_type:
+        section = file.file_type.split("/")[0]
 
     meat_f = f"""
     <meta property="og:type" content="article">
     <meta property="og:title" content="{file.name}">
     <meta property="og:url" content="{url}">
     <meta property="og:image" content="">
-    <meta property="og:description" content="{descr}">
+    <meta property="og:description" content="{html.escape(descr)}">
     <meta property="article:author" content="{file.user.username}">
+    <meta property="article:section" content="{section}">
     <meta property="article:published_time" content="{file.created}">
     <meta property="article:modified_time" content="{file.modified}">
     """

akarpov/files/previews/text/html.py

@@ -0,0 +1,82 @@
+import html
+
+from akarpov.files.models import File
+
+
+def view(file: File) -> (str, str):
+    static = f"""
+    <meta property="og:title" content="{file.name}" />
+<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/styles/atom-one-light.min.css">
+    """
+    content = "<div class='col-auto'><pre>"
+    with file.file.open("r") as f:
+        lines = f.readlines()
+    style = False
+    js = False
+    for line in lines:
+        if "<style" in line and "</style>" not in line:
+            style = True
+        elif "</style>" in line:
+            style = False
+
+        if "<script" in line and "</script>" not in line:
+            js = True
+        elif "</script>" in line:
+            js = False
+
+        if style:
+            content += f"""<div class='code language-css'>{html.escape(line)}</div>"""
+        elif js:
+            content += (
+                f"""<div class='code language-javascript'>{html.escape(line)}</div>"""
+            )
+        else:
+            content += f"""<div class='code language-html'>{html.escape(line)}</div>"""
+
+    content += """</pre></div>
+      <script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/highlight.min.js"></script>
+      <script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/languages/javascipt.min.js"></script>
+      <script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/languages/html.min.js"></script>
+      <script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/languages/css.min.js"></script>
+      <script>
+      hljs.configure({ ignoreUnescapedHTML: true })
+      document.querySelectorAll('div.code').forEach(el => {
+          hljs.highlightElement(el);
+        });
+      </script>
+    """
+
+    return static, content
+
+
+def meta(file: File):
+    descr = ""
+    i = 0
+    with file.file.open("r") as f:
+        lines = f.readlines()
+    for line in lines:
+        if i == 0:
+            descr += line + "\n"
+        else:
+            descr += line + "    "
+        i += 1
+        if i > 20:
+            descr += "..."
+            break
+    url = file.get_absolute_url()
+    section = ""
+    if file.file_type:
+        section = file.file_type.split("/")[0]
+
+    meat_f = f"""
+    <meta property="og:type" content="article">
+    <meta property="og:title" content="{file.name}">
+    <meta property="og:url" content="{url}">
+    <meta property="og:image" content="">
+    <meta property="og:description" content="{html.escape(descr)}">
+    <meta property="article:author" content="{file.user.username}">
+    <meta property="article:section" content="{section}">
+    <meta property="article:published_time" content="{file.created}">
+    <meta property="article:modified_time" content="{file.modified}">
+    """
+    return meat_f

akarpov/files/services/preview.py

@@ -4,6 +4,8 @@
 from PIL import Image, ImageDraw, ImageFont
 from preview_generator.manager import PreviewManager
 
+from akarpov.files.models import File
+
 cache_path = "/tmp/preview_cache"
 PIL_GRAYSCALE = "L"
 PIL_WIDTH_INDEX = 0
@@ -92,3 +94,26 @@ def get_description(file_path: str) -> str:
     if manager.has_text_preview(file_path):
         return manager.get_text_preview(file_path)
     return ""
+
+
+def get_base_meta(file: File):
+    preview = file.preview.url if file.preview else ""
+    description = file.description if file.description else ""
+    return f"""
+    <meta property="og:type" content="website">
+    <meta property="og:title" content="{file.name}">
+    <meta property="og:url" content="{file.get_absolute_url()}">
+    <meta property="og:image" content="{preview}">
+    <meta property="og:image:width" content="500" />
+    <meta property="og:image:height" content="500" />
+    <meta property="og:description" content="{description}">
+
+    <meta name="twitter:card" content="summary_large_image">
+    <meta name="twitter:title" content="{file.name}">
+    <meta name="twitter:site" content="{file.get_absolute_url()}">
+    <meta name="twitter:description" content="{description}">
+    <meta name="twitter:image" content="{preview}">
+    <meta property="twitter:image:width" content="500" />
+    <meta property="twitter:image:height" content="500" />
+    <meta name="twitter:image:alt" content="{file.name}">
+    """

akarpov/files/views.py

@@ -18,6 +18,7 @@
 from akarpov.files.forms import FileForm
 from akarpov.files.models import File, Folder
 from akarpov.files.previews import extensions, meta, meta_extensions, previews
+from akarpov.files.services.preview import get_base_meta
 
 logger = structlog.get_logger(__name__)
 
@@ -69,38 +70,47 @@ def dispatch(self, request, *args, **kwargs):
         return super().dispatch(request, *args, **kwargs)
 
     def get_context_data(self, **kwargs):
-        context = super().get_context_data(**kwargs)
-        context["has_perm"] = self.object.user == self.request.user
         static = ""
         content = ""
         meta_s = ""
         extension = self.object.file.path.split(".")[-1]
+
         try:
+            meta_loaded = False
             if self.object.file_type:
                 t1, t2 = self.object.file_type.split("/")
                 loaded = False
+
+                # get static and content for file view by mimetype
                 if t1 in previews:
                     if t2 in previews[t1]:
                         static, content = previews[t1][t2](self.object)
                         loaded = True
                 if not loaded and extension in extensions:
                     static, content = extensions[extension](self.object)
-            elif extension in extensions:
-                static, content = extensions[extension](self.object)
 
-            if self.object.file_type:
+                # get meta tags by mimetype
-                t1, t2 = self.object.file_type.split("/")
-                loaded = False
                 if t1 in meta:
                     if t2 in meta[t1]:
                         meta_s = meta[t1][t2](self.object)
-                        loaded = True
+                        meta_loaded = True
-                if not loaded and extension in meta_extensions:
+                if not meta_loaded and extension in meta_extensions:
                     meta_s = meta_extensions[extension](self.object)
-            elif extension in meta_extensions:
+            else:
-                meta_s = meta_extensions[extension](self.object)
+                # get static and content for file view by file extension
+                if extension in extensions:
+                    static, content = extensions[extension](self.object)
+                # get meta tags by file extension
+                if extension in meta_extensions:
+                    meta_s = meta_extensions[extension](self.object)
+                    meta_loaded = True
+            if not meta_loaded:
+                meta_s = get_base_meta(self.object)
         except Exception as e:
             logger.error(e)
+
+        context = super().get_context_data(**kwargs)
+        context["has_perm"] = self.object.user == self.request.user
         context["preview_static"] = static
         context["preview_content"] = content
         context["meta"] = meta_s

akarpov/users/signals.py

@@ -0,0 +1,11 @@
+from django.db.models.signals import pre_save
+from django.dispatch import receiver
+
+from akarpov.users.models import User
+
+
+@receiver(pre_save, sender=User)
+def user_create(sender, instance: User, **kwargs):
+    if instance.id is None:
+        # give user some space on file share on register
+        instance.left_file_upload += 100 * 1024 * 1024