added ml better support, better site notifications

akarpov/common/ml/text.py

@@ -0,0 +1,51 @@
+import pycld2 as cld2
+import spacy
+import torch
+from transformers import AutoModel, AutoTokenizer
+
+# load ml classes and models on first request
+# TODO: move to outer server/service
+nlp = None
+ru_nlp = None
+
+ru_model = None
+ru_tokenizer = None
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+
+def get_text_embedding(text: str):
+    global nlp, ru_nlp, ru_model, ru_tokenizer
+
+    is_reliable, text_bytes_found, details = cld2.detect(text)
+    if is_reliable:
+        lang = details[0]
+        if lang[1] in ["ru", "en"]:
+            lang = lang[1]
+        else:
+            return None
+    else:
+        return None
+
+    if lang == "ru":
+        if not ru_nlp:
+            ru_nlp = spacy.load("ru_core_news_md", disable=["parser", "ner"])
+        lema = " ".join([token.lemma_ for token in ru_nlp(text)])
+        if not ru_model:
+            ru_model = AutoModel.from_pretrained("DeepPavlov/rubert-base-cased")
+        if not ru_tokenizer:
+            ru_tokenizer = AutoTokenizer.from_pretrained("DeepPavlov/rubert-base-cased")
+        encodings = ru_tokenizer(
+            lema,  # the texts to be tokenized
+            padding=True,  # pad the texts to the maximum length (so that all outputs have the same length)
+            return_tensors="pt",  # return the tensors (not lists)
+        )
+        with torch.no_grad():
+            # get the model embeddings
+            embeds = ru_model(**encodings)
+        embeds = embeds[0]
+    elif lang == "en":
+        embeds = None
+    else:
+        embeds = None
+
+    return embeds

akarpov/files/migrations/0025_create_vector_ps.py

@@ -0,0 +1,10 @@
+from django.db import migrations
+from pgvector.django import VectorExtension
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("files", "0024_alter_file_options_alter_filereport_options_and_more"),
+    ]
+
+    operations = [VectorExtension()]

akarpov/files/migrations/0026_file_embeddings.py

@@ -0,0 +1,18 @@
+# Generated by Django 4.2.5 on 2023-09-16 18:33
+
+from django.db import migrations
+import pgvector.django
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("files", "0025_create_vector_ps"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="file",
+            name="embeddings",
+            field=pgvector.django.VectorField(dimensions=768, null=True),
+        ),
+    ]

akarpov/files/migrations/0027_merge_20230925_2023.py

@@ -0,0 +1,12 @@
+# Generated by Django 4.2.5 on 2023-09-25 17:23
+
+from django.db import migrations
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("files", "0025_file_notify_user_on_view_alter_basefileitem_parent"),
+        ("files", "0026_file_embeddings"),
+    ]
+
+    operations = []

akarpov/files/migrations/0028_file_content_file_lang.py

@@ -0,0 +1,26 @@
+# Generated by Django 4.2.5 on 2023-09-26 09:04
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ("files", "0027_merge_20230925_2023"),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name="file",
+            name="content",
+            field=models.TextField(default="", max_length=10000),
+            preserve_default=False,
+        ),
+        migrations.AddField(
+            model_name="file",
+            name="lang",
+            field=models.CharField(
+                choices=[("ru", "ru"), ("en", "en")], default="en", max_length=2
+            ),
+            preserve_default=False,
+        ),
+    ]

akarpov/files/models.py

@@ -17,6 +17,7 @@
 from django.urls import reverse
 from model_utils.fields import AutoCreatedField, AutoLastModifiedField
 from model_utils.models import TimeStampedModel
+from pgvector.django import VectorField
 from polymorphic.models import PolymorphicModel
 
 from akarpov.files.services.files import trash_file_upload, user_unique_file_upload
@@ -69,6 +70,9 @@ class File(BaseFileItem, TimeStampedModel, ShortLinkModel, UserHistoryModel):
 
     preview = FileField(blank=True, upload_to="file/previews/")
     file_obj = FileField(blank=False, upload_to=user_unique_file_upload)
+    embeddings = VectorField(dimensions=768, null=True)
+    content = TextField(max_length=10000)
+    lang = CharField(max_length=2, choices=[("ru", "ru"), ("en", "en")])
 
     # meta
     name = CharField(max_length=255, null=True, blank=True)

akarpov/files/services/text.py

@@ -0,0 +1,7 @@
+import textract
+
+
+def extract_file_text(file: str) -> str:
+    text = textract.process(file)
+
+    return text

akarpov/templates/base.html

@@ -132,6 +132,10 @@
                 <footer class="row bg-light py-1 mt-auto text-center">
                     <div class="col"> Writen by <a href="/about">sanspie</a>, find source code <a href="https://github.com/Alexander-D-Karpov/akarpov">here</a> </div>
                 </footer>
+            <div id="toastContainer" class="toast-container position-fixed bottom-0 end-0 p-3">
+
+            </div>
+
             </div>
         </div>
     </div>
@@ -142,18 +146,63 @@
     {% endblock inline_javascript %}
     {% if request.user.is_authenticated %}
     <script>
-      {# TODO: add automatic socket host retrieve  #}
-      let socket = new WebSocket(`ws://127.0.0.1:8000/ws/notifications/`);
+      {% if request.is_secure %}
+      let socket = new WebSocket(`wss://{{ request.get_host }}/ws/notifications/`);
+      {% else %}
+      let socket = new WebSocket(`ws://{{ request.get_host }}/ws/notifications/`);
+      {% endif %}
 
       function sleep(ms) {
           return new Promise(resolve => setTimeout(resolve, ms));
       }
 
+      function timeSince(date) {
+            let seconds = Math.floor((new Date() - date) / 1000);
+            let interval = seconds / 31536000;
+            if (interval > 1) {
+              return Math.floor(interval) + " years";
+            }
+            interval = seconds / 2592000;
+            if (interval > 1) {
+              return Math.floor(interval) + " months";
+            }
+            interval = seconds / 86400;
+            if (interval > 1) {
+              return Math.floor(interval) + " days";
+            }
+            interval = seconds / 3600;
+            if (interval > 1) {
+              return Math.floor(interval) + " hours";
+            }
+            interval = seconds / 60;
+            if (interval > 1) {
+              return Math.floor(interval) + " minutes";
+            }
+            return Math.floor(seconds) + " seconds";
+      }
+
+      const toastContainer = document.getElementById('toastContainer')
+
+
       let fn = async function(event) {
         let data = JSON.parse(event.data)
-        console.log(data)
-        alert(data.body)
-        {# TODO add pretty pop up #}
+        const toast = document.createElement("div")
+        toast.id = "liveToast"
+        toast.className = "toast mb-4 ml-2"
+        toast.setAttribute("role", "alert")
+        toast.setAttribute("aria-live", "assertive")
+        toast.setAttribute("aria-atomic", "true")
+        toast.innerHTML = `<div class="toast-header">
+                    <strong class="me-auto">${data.title}</strong>
+                    <small>${timeSince(Date.parse(data.created))} ago</small>
+                    <button type="button" class="btn-close" data-bs-dismiss="toast" aria-label="Close"></button>
+                  </div>
+                  <div class="toast-body">
+                    ${data.body}
+                  </div>`
+        toastContainer.appendChild(toast)
+        const toastBootstrap = bootstrap.Toast.getOrCreateInstance(toast)
+        toastBootstrap.show()
       }
 
       socket.onmessage = fn

compose/local/django/Dockerfile

@@ -28,6 +28,8 @@ RUN apt-get update && \
     apt-get install -y build-essential libpq-dev gettext libmagic-dev libjpeg-dev zlib1g-dev  && \
     # Dependencies for file preview generation
     apt-get install -y webp libimage-exiftool-perl libmagickwand-dev ffmpeg libgdal-dev && \
+    # ML dependencies \
+    # none for now
     apt-get purge -y --auto-remove -o APT:AutoRemove:RecommendsImportant=false && \
     rm -rf /var/lib/apt/lists/*
 

compose/production/postgres/Dockerfile

@@ -1,5 +1,10 @@
 FROM postgres:14
 
+RUN apt-get update && \
+    apt-get install -y postgresql-14-pgvector && \
+    apt-get purge -y --auto-remove -o APT:AutoRemove:RecommendsImportant=false && \
+    rm -rf /var/lib/apt/lists/*
+
 COPY ./compose/production/postgres/maintenance /usr/local/bin/maintenance
 RUN chmod +x /usr/local/bin/maintenance/*
 RUN mv /usr/local/bin/maintenance/* /usr/local/bin \

config/settings/local.py

@@ -61,5 +61,5 @@
 
 # SHORTENER
 # ------------------------------------------------------------------------------
-SHORTENER_REDIRECT_TO = "http://127.0.0.1:8000"
-SHORTENER_HOST = "http://127.0.0.1:3000"
+SHORTENER_REDIRECT_TO = "https://dev2.akarpov.ru"
+SHORTENER_HOST = "https://dev.akarpov.ru"

pyproject.toml

@@ -107,6 +107,8 @@ pytest-xdist = "^3.3.1"
 pytest-mock = "^3.11.1"
 pytest-asyncio = "^0.21.1"
 pytest-lambda = "^2.2.0"
+pgvector = "^0.2.2"
+pycld2 = "^0.41"
 
 
 [build-system]