MongoDB Atlas Vector Database: Zero to Advanced with Python

Connect to MongoDB using mongosh and Compass to manage databases and collections. Create and switch databases, add collections, and drop databases as shown.

Explore how MongoDB handles diverse data types in documents, including int, string, boolean, date, null, arrays, and an address document, with practical schema creation and insertion examples.

Explore MongoDB sorting and limiting data with find queries, using sort criteria plus one for ascending and minus one for descending, and applying limit to order by age or name.

Explore how to combine query conditions with logical operators in MongoDB, using and, or, not, and nor with age and salary to filter results.

Demonstrate update many in MongoDB to update multiple records for Sita, Lakshman, and Arjun by name, setting full time to true, then add a gift field for full time employees.

Learn how MongoDB Atlas, a fully managed cloud database, simplifies deployment and scaling. Create a free account, set up a cluster, and compare Atlas with the compass.

Install Python, set up a virtual environment, and install PyMongo to connect to MongoDB, then perform CRUD operations with PyMongo and verify the setup.

Connect your MongoDB Atlas cluster to a Python app with PyMongo, using the atlas connection string and a username and password. Verify the connection with a ping command.

Use PyMongo to print all documents by iterating over collection.find results and applying pretty print to display fields like ID, name, age, and salary.

Use PyMongo's insert_many to add multiple documents into a MongoDB collection, building a list of documents with fields such as name, age, and address. Verify by counting with find.

Master PyMongo's limit, sort, and skip to control query results, using ascending and descending order on fields like age and salary.

Explore how to use PyMongo's find operator for querying and projection, applying comparison operators such as $eq, and shaping results with projection controls.

Master PyMongo logical operators to combine conditions using $and and $or, querying age greater than 30 and salary at least 50,000, with name and salary projection.

Learn why rerunning PyMongo scripts can create duplicate records and how to prevent it by avoiding running all, dropping the database, and focusing on final queries.

Explore PyMongo's element query operators in MongoDB, focusing on $exists and $type to filter documents by field presence and data type, with practical examples.

Explore how to use PyMongo's array query operator $all to find documents whose array fields contain all specified elements, irrespective of order, with practical examples.

Explore how to query MongoDB arrays with PyMongo using $elemMatch to match documents where an array’s element meets multiple criteria, and use $size to filter arrays by length.

Learn how aggregation operations use pipelines of stages to group data, compute minimum, maximum, and average values, and pass results through successive stages to produce final results.

Learn to use the aggregate stage and the $match operator in PyMongo to filter documents in a pipeline, such as selecting records where sex is female.

Filter documents where sex is female and total bill is greater than $20 by applying the match operator in an aggregate pipeline using the and operator.

Explore the MongoDB aggregate group stage, using $group to cluster documents by a group key and compute totals with accumulators, producing one document per unique key in a pipeline.

demonstrates how aggregate stage $count tallies documents after a match, using a pipeline with tip amount >= 3 and an alias like 'count of documents' to return the total.

Explore MongoDB aggregation with PyMongo by using the $project stage to display employee id, total bill, and tip while excluding _id, and to create a concatenated day and time field.

Create a tip bucket in MongoDB's aggregation pipeline using the condition operator to implement an if-then-else that marks tips as high or low by value.

Learn the second syntax for the MongoDB $cond operator, using a boolean expression with a true and false case in a list to drive if-then-else logic.

Learn to combine project and group stages in a MongoDB aggregation pipeline to count documents by bill bucket values (high, medium, low) using a sum accumulator.

Discover how to use the addFields stage in MongoDB aggregation to add new fields, compute totals like total spend, and create conditional buckets such as tip bucket.

Explore linear interpolation to estimate unknown values between data points using a straight line, the slope formula m = (y1−y0)/(x1−x0), and y = y0 + m(x−x0) in MongoDB.

Learn how last observation carry forward (locf) uses forward fill and backward fill to estimate missing values, with sorting documents and filling values in a MongoDB workflow using PyMongo.

Learn to fill missing average IQ score using forward fill and backward fill with the $fill stage, sorting by a numerical field to carry the last observation forward or backward.

Explore how to fill missing values in MongoDB using the $fill stage, with partition by and sort by, applying forward fill, backward fill, or linear interpolation.

Explore the $merge stage, which writes aggregation results to a specified collection or database, with on-field matching and options to insert, replace, merge, discard, or fail when matched or not.

Learn to use MongoDB's aggregate $merge with PyMongo to merge documents into tips copy collection in my save data database, and create a tip bucket field bucketing by tip values.

Demonstrates the merge operator in the aggregation stage, covering when matched and when not matched behaviors, including merge, replace, keep existing, insert, discard, and fail options.

Explore how MongoDB's $text full text search uses NLP techniques like tokenization, normalization, diacritic sensitivity, case sensitivity, stemming, and stop-word removal to score and rank documents with language options.

Explore how to compute similarity measures, dot product, cosine similarity, Euclidean distance, and Manhattan distance, between vectors like king, queen, dog, and puppy, to identify the closest vector.

Learn to perform MongoDB vector search using exact and approximate nearest neighbor methods, leveraging cosine similarity, dot product, and Euclidean distance in aggregated pipelines with index search models.

Demonstrates setting up a MongoDB Atlas vector search with toy movie data, creating eight-dimensional embeddings, building a vector index, and querying with cosine similarity via an aggregate pipeline.

Set the OpenAI API key and Atlas connection string as environment variables using bash exports on Linux/macOS and set X on Windows, then verify with echo.

Explore the LangChain chat OpenAI model by building a chat interface with user, assistant, and system roles, and crafting prompts using system and human messages.

Master retrieval augmented generation with Langchain by mastering indexing, retrieving, and pre-processing documents into chunks and embeddings stored in MongoDB for effective llm insights.

Connect to MongoDB Atlas via a Python notebook, create MongoDB book database and collection, and set up Atlas Vector Search index with a 512-d text embeddings field and cosine similarity.

Create 94 document chunks from 66 documents using a recursive character text splitter with a chunk size of 1000 and 20% overlap. This prepares for generating embeddings from the chunks.

Learn how to store document chunks and their embeddings in a MongoDB Atlas vector search collection using long chain's from_documents, configure the embedding model, and verify indexing.

Master vector search on embeddings in MongoDB Atlas by building a two-stage aggregate pipeline, generating a query embedding, and projecting text, page, and similarity scores.

Learn to perform vector search in MongoDB Atlas using Langchain by directly querying the object vector store, performing similarity search with score, and retrieving metadata and page content.