celery beat + image embedding

db/migrations/versions/20250427_171537_initial_structure.py

@@ -361,13 +361,13 @@ def downgrade() -> None:
     # Drop tables
     op.drop_table("photo")
     op.drop_table("misc_doc")
-    op.drop_table("mail_message")
     op.drop_table("git_commit")
     op.drop_table("chat_message")
     op.drop_table("book_doc")
     op.drop_table("blog_post")
     op.drop_table("github_item")
     op.drop_table("email_attachment")
+    op.drop_table("mail_message")
     op.drop_table("source_item")
     op.drop_table("rss_feeds")
     op.drop_table("email_accounts")

docker-compose.yaml

@@ -21,6 +21,8 @@ volumes:
 # ------------------------------ X-templates ----------------------------
 x-common-env: &env
   RABBITMQ_USER: kb
+  QDRANT_HOST: qdrant
+  DB_HOST: postgres
   FILE_STORAGE_DIR: /app/memory_files
   TZ: "Etc/UTC"
 
@@ -165,6 +167,13 @@ services:
       # - ./acme.json:/acme.json:rw
 
 # ------------------------------------------------------------ Celery workers
+  worker-email:
+    <<: *worker-base
+    environment:
+      <<: *worker-env
+      QUEUES: "email"
+    deploy: {resources: {limits: {cpus: "2", memory: 3g}}}
+
   worker-text:
     <<: *worker-base
     environment:
@@ -207,6 +216,22 @@ services:
       QUEUES: "docs"
     deploy: {resources: {limits: {cpus: "1", memory: 1g}}}
 
+  ingest-hub:
+    <<: *worker-base
+    build: 
+      context: .
+      dockerfile: docker/ingest_hub/Dockerfile
+    environment:
+      <<: *worker-env
+    volumes:
+      - file_storage:/app/memory_files:rw
+    tmpfs:
+      - /tmp
+      - /var/tmp
+      - /var/log/supervisor
+      - /var/run/supervisor
+    deploy: {resources: {limits: {cpus: "0.5", memory: 512m}}}
+
 # ------------------------------------------------------------ watchtower (auto-update)
   watchtower:
     image: containrrr/watchtower

docker/ingest_hub/Dockerfile

@@ -0,0 +1,32 @@
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# Copy requirements files and setup
+COPY requirements-*.txt ./
+COPY setup.py ./
+COPY src/ ./src/
+
+# Install dependencies
+RUN apt-get update && apt-get install -y \
+        libpq-dev gcc supervisor && \
+    pip install -e ".[workers]" && \
+    apt-get purge -y gcc && apt-get autoremove -y && rm -rf /var/lib/apt/lists/*
+
+# Create and copy entrypoint script
+COPY docker/ingest_hub/supervisor.conf /etc/supervisor/conf.d/supervisor.conf
+COPY docker/workers/entry.sh ./entry.sh
+RUN chmod +x entry.sh
+
+# Create required tmpfs directories for supervisor
+RUN mkdir -p /var/log/supervisor /var/run/supervisor
+
+# Create user and set permissions
+RUN useradd -m kb && chown -R kb /app /var/log/supervisor /var/run/supervisor
+USER kb
+
+# Default queues to process
+ENV QUEUES="medium_embed,photo_embed,low_ocr,git_summary,rss,docs,email"
+ENV PYTHONPATH="/app"
+
+ENTRYPOINT ["/usr/bin/supervisord", "-c", "/etc/supervisor/conf.d/supervisor.conf"]

docker/ingest_hub/supervisor.conf

@@ -0,0 +1,16 @@
+[supervisord]
+nodaemon=true
+loglevel=info
+logfile=/dev/stdout
+logfile_maxbytes=0
+user=kb
+pidfile=/dev/null
+
+[program:celery-beat]
+command=celery -A memory.workers.ingest beat --pidfile= --schedule=/tmp/celerybeat-schedule
+stdout_logfile=/dev/stdout
+stdout_logfile_maxbytes=0
+stderr_logfile=/dev/stderr
+stderr_logfile_maxbytes=0
+autorestart=true
+startsecs=10

docker/workers/Dockerfile

@@ -22,7 +22,7 @@ RUN useradd -m kb && chown -R kb /app
 USER kb
 
 # Default queues to process
-ENV QUEUES="medium_embed,photo_embed,low_ocr,git_summary,rss,docs"
+ENV QUEUES="medium_embed,photo_embed,low_ocr,git_summary,rss,docs,email"
 ENV PYTHONPATH="/app"
 
 ENTRYPOINT ["./entry.sh"] 

requirements-common.txt

@@ -2,6 +2,7 @@ sqlalchemy==2.0.30
 psycopg2-binary==2.9.9
 pydantic==2.7.1
 alembic==1.13.1 
-dotenv==1.1.0
+dotenv==0.9.9
 voyageai==0.3.2
-qdrant-client==1.9.0 
+qdrant-client==1.9.0 
+PyMuPDF==1.25.5

src/memory/common/embedding.py

@@ -1,7 +1,9 @@
 import pathlib
-from typing import Literal, TypedDict, Iterable
+from typing import Literal, TypedDict, Iterable, Any
 import voyageai
 import re
+import uuid
+from memory.common import extract, settings
 
 # Chunking configuration
 MAX_TOKENS = 32000  # VoyageAI max context window
@@ -11,6 +13,8 @@ CHARS_PER_TOKEN = 4
 
 DistanceType = Literal["Cosine", "Dot", "Euclidean"]
 Vector = list[float]
+Embedding = tuple[str, Vector, dict[str, Any]]
+
 
 class Collection(TypedDict):
     dimension: int
@@ -163,17 +167,48 @@ def chunk_text(text: str, max_tokens: int = MAX_TOKENS, overlap: int = OVERLAP_T
         yield current.strip()
 
 
-def embed_text(text: str, model: str = "voyage-3-large", n_dimensions: int = 1536) -> list[Vector]:
+def embed_chunks(chunks: list[extract.MulitmodalChunk], model: str = settings.TEXT_EMBEDDING_MODEL) -> list[Vector]:
     vo = voyageai.Client()
-    return vo.embed(chunk_text(text, MAX_TOKENS, OVERLAP_TOKENS), model=model)
+    return vo.embed(chunks, model=model).embeddings
 
 
-def embed_file(file_path: pathlib.Path, model: str = "voyage-3-large", n_dimensions: int = 1536) -> list[Vector]:
-    return embed_text(file_path.read_text(), model, n_dimensions)
+def embed_text(texts: list[str], model: str = settings.TEXT_EMBEDDING_MODEL) -> list[Vector]:
+    chunks = [c for text in texts for c in chunk_text(text, MAX_TOKENS, OVERLAP_TOKENS) if c.strip()]
+    return embed_chunks(chunks, model)
 
 
-def embed(mime_type: str, content: bytes | str, model: str = "voyage-3-large", n_dimensions: int = 1536) -> tuple[str, list[Vector]]:
-    if isinstance(content, bytes):
-        content = content.decode("utf-8")
+def embed_file(file_path: pathlib.Path, model: str = settings.TEXT_EMBEDDING_MODEL) -> list[Vector]:
+    return embed_text([file_path.read_text()], model)
 
-    return get_modality(mime_type), embed_text(content, model, n_dimensions)
+
+def embed_mixed(items: list[extract.MulitmodalChunk], model: str = settings.MIXED_EMBEDDING_MODEL) -> list[Vector]:
+    def to_chunks(item: extract.MulitmodalChunk) -> Iterable[str]:
+        if isinstance(item, str):
+            return [c for c in chunk_text(item, MAX_TOKENS, OVERLAP_TOKENS) if c.strip()]
+        return [item]
+
+    chunks = [c for item in items for c in to_chunks(item)]
+    return embed_chunks(chunks, model)
+
+
+def embed_page(page: dict[str, Any]) -> list[Vector]:
+    contents = page["contents"]
+    if all(isinstance(c, str) for c in contents):
+        return embed_text(contents, model=settings.TEXT_EMBEDDING_MODEL)
+    return embed_mixed(contents, model=settings.MIXED_EMBEDDING_MODEL)
+
+
+def embed(
+    mime_type: str,
+    content: bytes | str | pathlib.Path,
+    metadata: dict[str, Any] = {},
+) -> tuple[str, list[Embedding]]:
+    modality = get_modality(mime_type)
+
+    pages = extract.extract_content(mime_type, content)
+    vectors = [
+        (str(uuid.uuid4()), vector, page.get("metadata", {}) | metadata)
+        for page in pages
+        for vector in embed_page(page)
+    ]
+    return modality, vectors

src/memory/common/extract.py

@@ -0,0 +1,74 @@
+from contextlib import contextmanager
+import io
+import pathlib
+import tempfile
+import pymupdf  # PyMuPDF
+from PIL import Image
+from typing import Any, TypedDict, Generator
+
+
+MulitmodalChunk = Image.Image | str
+class Page(TypedDict):
+    contents: list[MulitmodalChunk]
+    metadata: dict[str, Any]
+
+
+@contextmanager
+def as_file(content: bytes | str | pathlib.Path) -> Generator[pathlib.Path, None, None]:
+    if isinstance(content, pathlib.Path):
+        yield content
+    else:
+        mode = "w" if isinstance(content, str) else "wb"
+        with tempfile.NamedTemporaryFile(mode=mode) as f:
+            f.write(content)
+            f.flush()
+            yield pathlib.Path(f.name)
+
+
+def page_to_image(page: pymupdf.Page) -> Image.Image:
+    pix = page.get_pixmap()
+    return Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
+
+
+def doc_to_images(content: bytes | str | pathlib.Path) -> list[Page]:
+    with as_file(content) as file_path:
+        with pymupdf.open(file_path) as pdf:
+            return [{
+                "contents": page_to_image(page),
+                "metadata": {
+                    "page": page.number,
+                    "width": page.rect.width,
+                    "height": page.rect.height,
+                }
+            } for page in pdf.pages()]
+
+
+def extract_image(content: bytes | str | pathlib.Path) -> list[Page]:
+    if isinstance(content, pathlib.Path):
+        image = Image.open(content)
+    elif isinstance(content, bytes):
+        image = Image.open(io.BytesIO(content))
+    else:
+        raise ValueError(f"Unsupported content type: {type(content)}")
+    return [{"contents": image, "metadata": {}}]
+
+
+def extract_text(content: bytes | str | pathlib.Path) -> list[Page]:
+    if isinstance(content, pathlib.Path):
+        content = content.read_text()
+    if isinstance(content, bytes):
+        content = content.decode("utf-8")
+
+    return [{"contents": [content], "metadata": {}}]
+
+
+def extract_content(mime_type: str, content: bytes | str | pathlib.Path) -> list[Page]:
+    if mime_type == "application/pdf":
+        return doc_to_images(content)
+    if mime_type.startswith("text/"):
+        return extract_text(content)
+    if mime_type.startswith("image/"):
+        return extract_image(content)
+    
+    # Return empty list for unknown mime types
+    return []

src/memory/common/settings.py

@@ -7,16 +7,19 @@ load_dotenv()
 def boolean_env(key: str, default: bool = False) -> bool:
     return os.getenv(key, "0").lower() in ("1", "true", "yes")
 
-
+# Database settings
 DB_USER = os.getenv("DB_USER", "kb")
-DB_PASSWORD = os.getenv("DB_PASSWORD", "kb")
+if password_file := os.getenv("POSTGRES_PASSWORD_FILE"):
+    DB_PASSWORD = pathlib.Path(password_file).read_text().strip()
+else:
+    DB_PASSWORD = os.getenv("DB_PASSWORD", "kb")
+
 DB_HOST = os.getenv("DB_HOST", "postgres")
 DB_PORT = os.getenv("DB_PORT", "5432")
 DB_NAME = os.getenv("DB_NAME", "kb")
 
 def make_db_url(user=DB_USER, password=DB_PASSWORD, host=DB_HOST, port=DB_PORT, db=DB_NAME):
     return f"postgresql://{user}:{password}@{host}:{port}/{db}"
-
 DB_URL = os.getenv("DATABASE_URL", make_db_url())
 
 
@@ -32,4 +35,14 @@ QDRANT_PORT = int(os.getenv("QDRANT_PORT", "6333"))
 QDRANT_GRPC_PORT = int(os.getenv("QDRANT_GRPC_PORT", "6334"))
 QDRANT_API_KEY = os.getenv("QDRANT_API_KEY", None)
 QDRANT_PREFER_GRPC = boolean_env("QDRANT_PREFER_GRPC", False)  
-QDRANT_TIMEOUT = int(os.getenv("QDRANT_TIMEOUT", "60"))
+QDRANT_TIMEOUT = int(os.getenv("QDRANT_TIMEOUT", "60"))
+
+
+# Worker settings
+BEAT_LOOP_INTERVAL = int(os.getenv("BEAT_LOOP_INTERVAL", 3600))
+EMAIL_SYNC_INTERVAL = int(os.getenv("EMAIL_SYNC_INTERVAL", 600))
+
+
+# Embedding settings
+TEXT_EMBEDDING_MODEL = os.getenv("TEXT_EMBEDDING_MODEL", "voyage-3-large")
+MIXED_EMBEDDING_MODEL = os.getenv("MIXED_EMBEDDING_MODEL", "voyage-multimodal-3")

src/memory/workers/celery_app.py

@@ -1,6 +1,6 @@
 import os
 from celery import Celery
-
+from memory.common import settings
 
 def rabbit_url() -> str:
     user = os.getenv("RABBITMQ_USER", "guest")
@@ -8,10 +8,11 @@ def rabbit_url() -> str:
     return f"amqp://{user}:{password}@rabbitmq:5672//"
 
 
-app = Celery("memory",
-             broker=rabbit_url(),
-             backend=os.getenv("CELERY_RESULT_BACKEND",
-                               "db+postgresql://kb:kb@postgres/kb"))
+app = Celery(
+    "memory",
+    broker=rabbit_url(),
+    backend=os.getenv("CELERY_RESULT_BACKEND", f"db+{settings.DB_URL}")
+)
 
 
 app.autodiscover_tasks(["memory.workers.tasks"])
@@ -22,8 +23,8 @@ app.conf.update(
     task_reject_on_worker_lost=True,
     worker_prefetch_multiplier=1,
     task_routes={
-        # Task routing configuration
         "memory.workers.tasks.text.*": {"queue": "medium_embed"},
+        "memory.workers.tasks.email.*": {"queue": "email"},
         "memory.workers.tasks.photo.*": {"queue": "photo_embed"},
         "memory.workers.tasks.ocr.*": {"queue": "low_ocr"},
         "memory.workers.tasks.git.*": {"queue": "git_summary"},

src/memory/workers/email.py

@@ -14,7 +14,7 @@ from sqlalchemy.orm import Session
 from collections import defaultdict
 from memory.common import settings, embedding
 from memory.common.db.models import EmailAccount, MailMessage, SourceItem, EmailAttachment
-from memory.common.qdrant import get_qdrant_client, upsert_vectors
+from memory.common import qdrant
 
 logger = logging.getLogger(__name__)
 
@@ -471,17 +471,18 @@ def imap_connection(account: EmailAccount) -> Generator[imaplib.IMAP4_SSL, None,
 
 
 def vectorize_email(email: MailMessage) -> list[float]:
-    qdrant_client = get_qdrant_client()
+    qdrant_client = qdrant.get_qdrant_client()
 
-    chunks = embedding.embed_text(email.body_raw)
-    payloads = [email.as_payload()] * len(chunks)
-    vector_ids = [str(uuid.uuid4()) for _ in chunks]
-    upsert_vectors(
+    _, chunks = embedding.embed(
+        "text/plain", email.body_raw, metadata=email.as_payload(),
+    )
+    vector_ids, vectors, metadata = zip(*chunks)
+    qdrant.upsert_vectors(
         client=qdrant_client,
         collection_name="mail",
         ids=vector_ids,
-        vectors=chunks,
-        payloads=payloads,
+        vectors=vectors,
+        payloads=metadata,
     )
     vector_ids = [f"mail/{vector_id}" for vector_id in vector_ids]
 
@@ -491,15 +492,14 @@ def vectorize_email(email: MailMessage) -> list[float]:
             content = pathlib.Path(attachment.file_path).read_bytes()
         else:
             content = attachment.content
-        collection, vectors = embedding.embed(attachment.content_type, content)
-        attachment.source.vector_ids = vector_ids
-        embeds[collection].extend(
-            (str(uuid.uuid4()), vector, attachment.as_payload()) for vector in vectors
-        )
+        collection, chunks = embedding.embed(attachment.content_type, content, metadata=attachment.as_payload())
+        ids, vectors, metadata = zip(*chunks)
+        attachment.source.vector_ids = ids
+        embeds[collection].extend(chunks)
 
-    for collection, embeds in embeds.items():
-        ids, vectors, payloads = zip(*embeds)
-        upsert_vectors(
+    for collection, chunks in embeds.items():
+        ids, vectors, payloads = zip(*chunks)
+        qdrant.upsert_vectors(
             client=qdrant_client,
             collection_name=collection,
             ids=ids,

src/memory/workers/ingest.py

@@ -0,0 +1,14 @@
+from celery.schedules import schedule
+from memory.workers.tasks.email import SYNC_ALL_ACCOUNTS
+from memory.common import settings
+from memory.workers.celery_app import app
+
+
+@app.on_after_configure.connect
+def register_mail_schedules(sender, **_):
+    sender.add_periodic_task(
+        schedule=schedule(settings.EMAIL_SYNC_INTERVAL),
+        sig=app.signature(SYNC_ALL_ACCOUNTS),
+        name="sync-mail-all",
+        options={"queue": "email"},
+    )

src/memory/workers/tasks/__init__.py

@@ -1,4 +1,8 @@
 """
 Import sub-modules so Celery can register their @app.task decorators.
 """
-from memory.workers.tasks import text, photo, ocr, git, rss, docs, email   # noqa
+from memory.workers.tasks import text, photo, ocr, git, rss, docs, email   # noqa
+from memory.workers.tasks.email import SYNC_ACCOUNT, SYNC_ALL_ACCOUNTS, PROCESS_EMAIL
+
+
+__all__ = ["text", "photo", "ocr", "git", "rss", "docs", "email", "SYNC_ACCOUNT", "SYNC_ALL_ACCOUNTS", "PROCESS_EMAIL"]

src/memory/workers/tasks/email.py

@@ -18,8 +18,12 @@ from memory.workers.email import (
 
 logger = logging.getLogger(__name__)
 
+PROCESS_EMAIL = "memory.workers.tasks.email.process_message"
+SYNC_ACCOUNT = "memory.workers.tasks.email.sync_account"
+SYNC_ALL_ACCOUNTS = "memory.workers.tasks.email.sync_all_accounts"
 
-@app.task(name="memory.email.process_message")
+
+@app.task(name=PROCESS_EMAIL)
 def process_message(
     account_id: int, message_id: str, folder: str, raw_email: str,
 ) -> int | None:
@@ -35,6 +39,7 @@ def process_message(
     Returns:
         source_id if successful, None otherwise
     """
+    logger.info(f"Processing message {message_id} for account {account_id}")
     if not raw_email.strip():
         logger.warning(f"Empty email message received for account {account_id}")
         return None
@@ -76,7 +81,7 @@ def process_message(
         return source_item.id
 
 
-@app.task(name="memory.email.sync_account")
+@app.task(name=SYNC_ACCOUNT)
 def sync_account(account_id: int) -> dict:
     """
     Synchronize emails from a specific account.
@@ -87,6 +92,7 @@ def sync_account(account_id: int) -> dict:
     Returns:
         dict with stats about the sync operation
     """
+    logger.info(f"Syncing account {account_id}")
     with make_session() as db:
         account = db.query(EmailAccount).filter(EmailAccount.id == account_id).first()
         if not account or not account.active:
@@ -124,7 +130,7 @@ def sync_account(account_id: int) -> dict:
         }
 
 
-@app.task(name="memory.email.sync_all_accounts")
+@app.task(name=SYNC_ALL_ACCOUNTS)
 def sync_all_accounts() -> list[dict]:
     """
     Synchronize all active email accounts.

tests/conftest.py

@@ -7,6 +7,7 @@ from pathlib import Path
 
 import pytest
 import qdrant_client
+import voyageai
 from sqlalchemy import create_engine, text
 from sqlalchemy.orm import sessionmaker
 from testcontainers.qdrant import QdrantContainer
@@ -210,3 +211,8 @@ def qdrant():
             initialize_collections(client)
             yield client
 
+
+@pytest.fixture(autouse=True)
+def mock_voyage_client():
+    with patch.object(voyageai, "Client", autospec=True) as mock_client:
+        yield mock_client()

tests/memory/common/test_embedding.py

@@ -1,5 +1,19 @@
+import uuid
 import pytest
-from memory.common.embedding import yield_word_chunks, yield_spans, chunk_text, CHARS_PER_TOKEN
+from unittest.mock import Mock, patch
+from memory.common.embedding import yield_word_chunks, yield_spans, chunk_text, CHARS_PER_TOKEN, approx_token_count, get_modality, embed_text, embed_file, embed_mixed, embed_page, embed
+
+
+@pytest.fixture
+def mock_embed(mock_voyage_client):
+    vectors = ([i] for i in range(1000))
+
+    def embed(texts, model):
+        return Mock(embeddings=[next(vectors) for _ in texts])
+
+    mock_voyage_client.embed.side_effect = embed
+
+    return mock_voyage_client
 
 
 @pytest.mark.parametrize(
@@ -270,3 +284,77 @@ def test_chunk_text_very_long_sentences():
         'chunks by the function.',
     ]
 
+
+@pytest.mark.parametrize(
+    "string, expected_count",
+    [
+        ("", 0),
+        ("a" * CHARS_PER_TOKEN, 1),
+        ("a" * (CHARS_PER_TOKEN * 2), 2),
+        ("a" * (CHARS_PER_TOKEN * 2 + 1), 2),  # Truncation
+        ("a" * (CHARS_PER_TOKEN - 1), 0),  # Truncation
+    ]
+)
+def test_approx_token_count(string, expected_count):
+    assert approx_token_count(string) == expected_count
+
+
+@pytest.mark.parametrize(
+    "mime_type, expected_modality",
+    [
+        ("text/plain", "doc"),
+        ("text/html", "doc"),
+        ("image/jpeg", "photo"),
+        ("image/png", "photo"),
+        ("application/pdf", "book"),
+        ("application/epub+zip", "book"),
+        ("application/mobi", "book"),
+        ("application/x-mobipocket-ebook", "book"),
+        ("audio/mp3", "unknown"),
+        ("video/mp4", "unknown"),
+        ("text/something-new", "doc"),  # Should match by 'text/' stem
+        ("image/something-new", "photo"),  # Should match by 'image/' stem
+        ("custom/format", "unknown"),  # No matching stem
+    ]
+)
+def test_get_modality(mime_type, expected_modality):
+    assert get_modality(mime_type) == expected_modality
+
+
+def test_embed_text(mock_embed):
+    texts = ["text1 with words", "text2"]
+    assert embed_text(texts) == [[0], [1]]
+
+
+def test_embed_file(mock_embed, tmp_path):
+    mock_file = tmp_path / "test.txt"
+    mock_file.write_text("file content")
+    
+    assert embed_file(mock_file) == [[0]]
+
+
+def test_embed_mixed(mock_embed):
+    items = ["text", {"type": "image", "data": "base64"}]
+    assert embed_mixed(items) == [[0], [1]]
+
+
+def test_embed_page_text_only(mock_embed):
+    page = {"contents": ["text1", "text2"]}
+    assert embed_page(page) == [[0], [1]]
+
+
+def test_embed_page_mixed_content(mock_embed):
+    page = {"contents": ["text", {"type": "image", "data": "base64"}]}
+    assert embed_page(page) == [[0], [1]]
+
+
+def test_embed(mock_embed):
+    mime_type = "text/plain"
+    content = "sample content"
+    metadata = {"source": "test"}
+    
+    with patch.object(uuid, "uuid4", return_value="id1"):
+        modality, vectors = embed(mime_type, content, metadata)
+    
+    assert modality == "doc"
+    assert vectors == [('id1', [0], {'source': 'test'})]

tests/memory/common/test_extract.py

@@ -0,0 +1,131 @@
+import pathlib
+import pytest
+import pymupdf
+from PIL import Image
+import io
+from memory.common.extract import as_file, extract_text, extract_content, Page, doc_to_images, extract_image
+
+
+REGULAMIN = pathlib.Path(__file__).parent.parent.parent / "data" / "regulamin.pdf"
+
+
+def test_as_file_with_path(tmp_path):
+    test_path = tmp_path / "test.txt"
+    test_path.write_text("test content")
+    
+    with as_file(test_path) as path:
+        assert path == test_path
+        assert path.read_text() == "test content"
+
+
+def test_as_file_with_bytes():
+    content = b"test content"
+    with as_file(content) as path:
+        assert pathlib.Path(path).read_bytes() == content
+
+
+def test_as_file_with_str():
+    content = "test content"
+    with as_file(content) as path:
+        assert pathlib.Path(path).read_text() == content
+
+
+@pytest.mark.parametrize(
+    "input_content,expected",
+    [
+        ("simple text", [{"contents": ["simple text"], "metadata": {}}]),
+        (b"bytes text", [{"contents": ["bytes text"], "metadata": {}}]),
+    ]
+)
+def test_extract_text(input_content, expected):
+    assert extract_text(input_content) == expected
+
+
+def test_extract_text_with_path(tmp_path):
+    test_file = tmp_path / "test.txt"
+    test_file.write_text("file text content")
+    
+    assert extract_text(test_file) == [{"contents": ["file text content"], "metadata": {}}]
+
+
+def test_doc_to_images():
+    result = doc_to_images(REGULAMIN)
+    
+    assert len(result) == 2
+    with pymupdf.open(REGULAMIN) as pdf:
+        for page, pdf_page in zip(result, pdf.pages()):
+            pix = pdf_page.get_pixmap()
+            img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
+            assert page["contents"] == img
+            assert page["metadata"] == {
+                "page": pdf_page.number,
+                "width": pdf_page.rect.width,
+                "height": pdf_page.rect.height,
+            }
+
+
+def test_extract_image_with_path(tmp_path):
+    img = Image.new('RGB', (100, 100), color='red')
+    img_path = tmp_path / "test.png"
+    img.save(img_path)
+
+    page, = extract_image(img_path)
+    assert page["contents"].tobytes() == img.convert("RGB").tobytes()
+    assert page["metadata"] == {}
+
+
+def test_extract_image_with_bytes():
+    img = Image.new('RGB', (100, 100), color='blue')
+    buffer = io.BytesIO()
+    img.save(buffer, format='PNG')
+    img_bytes = buffer.getvalue()
+    
+    page, = extract_image(img_bytes)
+    assert page["contents"].tobytes() == img.convert("RGB").tobytes()
+    assert page["metadata"] == {}
+
+
+def test_extract_image_with_str():
+    with pytest.raises(ValueError):
+        extract_image("test")
+
+
+@pytest.mark.parametrize(
+    "mime_type,content",
+    [
+        ("text/plain", "Text content"),
+        ("text/html", "<html>content</html>"),
+        ("text/markdown", "# Heading"),
+        ("text/csv", "a,b,c"),
+    ]
+)
+def test_extract_content_different_text_types(mime_type, content):
+    assert extract_content(mime_type, content) == [{"contents": [content], "metadata": {}}]
+
+
+def test_extract_content_pdf():
+    result = extract_content("application/pdf", REGULAMIN)
+    
+    assert len(result) == 2
+    assert all(isinstance(page["contents"], Image.Image) for page in result)
+    assert all("page" in page["metadata"] for page in result)
+    assert all("width" in page["metadata"] for page in result)
+    assert all("height" in page["metadata"] for page in result)
+
+
+def test_extract_content_image(tmp_path):
+    # Create a test image
+    img = Image.new('RGB', (100, 100), color='red')
+    img_path = tmp_path / "test_img.png"
+    img.save(img_path)
+    
+    result = extract_content("image/png", img_path)
+    
+    assert len(result) == 1
+    assert isinstance(result[0]["contents"], Image.Image)
+    assert result[0]["contents"].size == (100, 100)
+    assert result[0]["metadata"] == {}
+
+
+def test_extract_content_unsupported_type():
+    assert extract_content("unsupported/type", "content") == []