WEBVTT 1 00:00:01.418 --> 00:00:07.674 In this session. we will implement a feature to log fire detection data into the database. 00:00:08.174 --> 00:00:11.803 Since this functionality involves handling the database. 00:00:12.178 --> 00:00:19.394 we will add it to the `crud` file within the `predict_fire` domain that we previously created. 00:00:20.020 --> 00:00:26.317 The feature itself is simple—process the detected data and store it in the database. 00:00:26.901 --> 00:00:31.489 The official documentation covers how to use SQLAlchemy. 00:00:31.823 --> 00:00:35.618 But since the content is extensive and not always intuitive. 00:00:35.785 --> 00:00:39.706 it is convenient to ask an LLM first when needed. 00:00:54.512 --> 00:00:58.933 Here. `add()`. `commit()`. and `refresh()` serve the following functions. 00:00:59.559 --> 00:01:05.774 `add()` adds an object to the session. The changes are only applied after calling `commit()`. 00:01:06.274 --> 00:01:09.319 `commit()` saves the changes to the database. 00:01:09.986 --> 00:01:14.908 `refresh()` fetches the latest data from the database and updates the object. 00:01:15.658 --> 00:01:22.373 To modify the database structure. we will delete the existing database file and recreate it. 00:01:23.708 --> 00:01:27.378 We considered adding a login feature. 00:01:27.378 --> 00:01:32.425 But including it in this lecture would unnecessarily increase the complexity. 00:01:32.425 --> 00:01:34.219 so we will not add it this time. 00:01:34.969 --> 00:01:41.851 Therefore. we will remove the relationship between the user table and the detection log table. 00:01:47.774 --> 00:01:55.615 We will change the result file key used for logging to the UUID-generated filename we created earlier. 00:01:56.199 --> 00:02:02.122 By storing this key. we can later access the image files saved in the log folder. 00:02:03.957 --> 00:02:12.882 We will call the `create_detection_log` function implemented in the `crud` file. To log data whenever a fire is detected. 00:02:16.886 --> 00:02:19.055 We will update the file path. 00:02:19.055 --> 00:02:24.435 And use FastAPI's `Depends` to inject a new DB session for each request. 00:02:26.855 --> 00:02:32.402 Once the request is completed. the session will be automatically cleaned up. 00:02:36.698 --> 00:02:42.787 We will run the server and check the database file using the SQLite extension. 00:02:58.595 --> 00:03:05.685 Using Postman. we will send a request to the `predict_fire` endpoint 00:03:11.482 --> 00:03:15.653 We can verify that the data has been successfully logged.