erioon 0.0.8__tar.gz → 0.0.9__tar.gz

{erioon-0.0.8 → erioon-0.0.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: erioon
-Version: 0.0.8
+Version: 0.0.9
 Summary: Erioon SDF for Python
 Author: Zyber Pireci
 Author-email: zyber.pireci@erioon.com

{erioon-0.0.8 → erioon-0.0.9}/erioon/collection.py

@@ -1,7 +1,7 @@
 import json
 from urllib.parse import urlparse
-from erioon.read import handle_get_all, handle_get_data
-from erioon.create import handle_insert_one, handle_insert_many
+from erioon.read import handle_get_all, handle_get_data, handle_classify_vector
+from erioon.create import handle_insert_one, handle_insert_many, handle_vector, handle_insert_many_vectors
 from erioon.delete import handle_delete_one, handle_delete_many
 from erioon.update import handle_update_query
 from erioon.ping import handle_connection_ping
@@ -223,6 +223,70 @@ class Collection:
             coll_id=self.coll_id,
             container_url=self.container_url,
         )
+        
+    def insert_one_vector(self, vector_data, metadata):
+        """
+        Insert a single record into the collection.
+
+        Args:
+            record (dict): Record to insert.
+
+        Returns:
+            tuple: (response message, HTTP status code)
+        """
+        if self._is_read_only():
+            return self._read_only_response()
+        return handle_vector(
+            user_id_cont=self.user_id,
+            database=self.db_id,
+            collection=self.coll_id,
+            vector=vector_data,
+            metadata=metadata,
+            container_url=self.container_url,
+        )
+        
+    def insert_many_vectors(self, records):
+        """
+        Insert multiple vector records into the collection.
+
+        Args:
+            records (list): List of dicts, each with keys 'vector', 'metadata', and optional '_id'.
+
+        Returns:
+            tuple: (response message, HTTP status code)
+        """
+        if self._is_read_only():
+            return self._read_only_response()
+
+        return handle_insert_many_vectors(
+            user_id_cont=self.user_id,
+            database=self.db_id,
+            collection=self.coll_id,
+            records=records,
+            container_url=self.container_url,
+        )
+        
+    def classify_vector(self, k=3):
+        """
+        Retrieve all vector records from the collection and classify them using k-NN.
+
+        Args:
+            k (int): Number of neighbors to use for classification.
+
+        Returns:
+            tuple: (response message, HTTP status code)
+        """
+        if self._is_read_only():
+            return self._read_only_response()
+            
+        return handle_classify_vector(
+            user_id=self.user_id,
+            db_id=self.db_id,
+            coll_id=self.coll_id,
+            container_url=self.container_url,
+            k=k
+        )
+
 
     def __str__(self):
         """Pretty print the collection metadata."""

{erioon-0.0.8 → erioon-0.0.9}/erioon/create.py

@@ -157,3 +157,109 @@ def handle_insert_many(user_id_cont, database, collection, data, container_url):
         general_error_msg = f"Unexpected error during bulk insert: {str(e)}"
         async_log(user_id_cont, database, collection, "POST", "ERROR", general_error_msg, 1, container_url)
         return {"status": "KO", "message": general_error_msg}, 500
+
+def handle_vector(user_id_cont, database, collection, vector, metadata, container_url, _id=None):
+    """
+    Inserts a vector embedding with its metadata into the blob-based vector DB.
+
+    Args:
+        user_id_cont: User identifier or context.
+        database: Database name.
+        collection: Collection name.
+        vector: List of floats representing the embedding.
+        metadata: Dictionary of associated metadata.
+        container_url: Azure Blob container SAS URL.
+        _id: Optional. Custom record ID. If None, UUID is generated.
+
+    Returns:
+        Tuple of (response dict, status code).
+    """
+    try:
+        if not isinstance(vector, list) or not all(isinstance(v, (float, int)) for v in vector):
+            return {"status": "KO", "message": "Invalid vector format. Must be a list of floats."}, 400
+
+        record = {
+            "_id": _id or str(uuid.uuid4()),
+            "vector": vector,
+            "metadata": metadata or {}
+        }
+
+        return handle_insert_one(user_id_cont, database, collection, record, container_url)
+
+    except Exception as e:
+        error_msg = f"Error in handle_vector: {str(e)}"
+        async_log(user_id_cont, database, collection, "POST", "ERROR", error_msg, 1, container_url)
+        return {"status": "KO", "message": "Failed to insert vector record.", "error": str(e)}, 500
+
+
+def handle_insert_many_vectors(user_id_cont, database, collection, records, container_url):
+    """
+    Bulk insert multiple vector embedding records with metadata into the blob-based vector DB.
+
+    Args:
+        user_id_cont: User identifier or context.
+        database: Database name.
+        collection: Collection name.
+        records: List of dicts, each with keys 'vector', 'metadata', and optional '_id'.
+        container_url: Azure Blob container SAS URL.
+
+    Returns:
+        Tuple of (response dict, status code).
+    """
+    insert_results = []
+
+    try:
+        for record in records:
+            vector = record.get("vector")
+            metadata = record.get("metadata", {})
+            _id = record.get("_id")
+
+            # Validate vector format
+            if not isinstance(vector, list) or not all(isinstance(v, (float, int)) for v in vector):
+                insert_results.append({
+                    "_id": _id,
+                    "status": "KO",
+                    "message": "Invalid vector format. Must be a list of floats."
+                })
+                continue
+
+            # Build record to insert
+            rec_to_insert = {
+                "_id": _id or str(uuid.uuid4()),
+                "vector": vector,
+                "metadata": metadata,
+            }
+
+            # Insert single vector record
+            response, status_code = handle_insert_one(
+                user_id_cont,
+                database,
+                collection,
+                rec_to_insert,
+                container_url
+            )
+
+            insert_results.append({
+                "_id": rec_to_insert["_id"],
+                "status": "OK" if status_code == 200 else "KO",
+                "message": response.get("message", "") if isinstance(response, dict) else str(response)
+            })
+
+        # Log bulk insert operation asynchronously
+        async_log(
+            user_id_cont,
+            database,
+            collection,
+            "POST",
+            "SUCCESS",
+            insert_results,
+            len(records),
+            container_url
+        )
+
+        return {"success": "Bulk vector insert completed.", "details": insert_results}, 200
+
+    except Exception as e:
+        error_msg = f"Unexpected error during bulk vector insert: {str(e)}"
+        async_log(user_id_cont, database, collection, "POST", "ERROR", error_msg, 1, container_url)
+        return {"status": "KO", "message": error_msg}, 500

{erioon-0.0.8 → erioon-0.0.9}/erioon/read.py

@@ -2,6 +2,7 @@ import io
 import msgpack
 from azure.storage.blob import ContainerClient
 from erioon.functions import async_log
+from sklearn.neighbors import KNeighborsClassifier
 
 
 def handle_get_all(user_id, db_id, coll_id, limit, container_url):
@@ -181,3 +182,60 @@ def handle_get_data(user_id, db_id, coll_id, search_criteria, limit, container_u
 
     async_log(user_id, db_id, coll_id, "GET", "ERROR", "No matching record found", 1, container_url)
     return {"status": "KO", "count": 0, "error": "No matching record found"}, 404
+
+
+def handle_classify_vector(user_id, db_id, coll_id, container_url, k=3):
+    # 1. Retrieve all data from collection
+    response, status = handle_get_all(user_id, db_id, coll_id, limit=10000, container_url=container_url)
+    if status != 200:
+        return {"status": "KO", "message": "Failed to fetch data for classification", "error": response.get("error", "")}, status
+
+    records = response.get("results", [])
+    if not records:
+        return {"status": "KO", "message": "No data found for classification"}, 404
+
+    # 2. Prepare dataset for classification
+    vectors = []
+    labels = []
+    unknown_vectors = []
+    unknown_ids = []
+
+    for rec in records:
+        vec = rec.get("vector")
+        meta = rec.get("metadata", {})
+        if not vec:
+            continue
+        vectors.append(vec)
+        labels.append(meta.get("class", "unknown"))
+    
+    # If some records don’t have classes, you might want to separate them or remove them
+    known_vectors = []
+    known_labels = []
+    for v, l in zip(vectors, labels):
+        if l != "unknown" and l is not None:
+            known_vectors.append(v)
+            known_labels.append(l)
+        else:
+            unknown_vectors.append(v)
+            unknown_ids.append(rec.get("_id"))
+
+    if not known_vectors:
+        return {"status": "KO", "message": "No labeled data for training classification"}, 404
+
+    # 3. Train k-NN classifier on known labeled vectors
+    knn = KNeighborsClassifier(n_neighbors=k)
+    knn.fit(known_vectors, known_labels)
+
+    # 4. Predict class for unknown vectors (if any)
+    predictions = {}
+    if unknown_vectors:
+        predicted_labels = knn.predict(unknown_vectors)
+        for _id, pred in zip(unknown_ids, predicted_labels):
+            predictions[_id] = pred
+
+    # 5. Return predictions (or full classification result)
+    return {
+        "status": "OK",
+        "message": f"Classification done on {len(unknown_vectors)} unknown vectors",
+        "predictions": predictions,
+    }, 200

{erioon-0.0.8 → erioon-0.0.9}/erioon.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: erioon
-Version: 0.0.8
+Version: 0.0.9
 Summary: Erioon SDF for Python
 Author: Zyber Pireci
 Author-email: zyber.pireci@erioon.com

{erioon-0.0.8 → erioon-0.0.9}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name='erioon',
-    version='0.0.8',
+    version='0.0.9',
     author='Zyber Pireci',
     author_email='zyber.pireci@erioon.com',
     description='Erioon SDF for Python',