SensorLogger Architecture

The sensor logger bridges sensors and file I/O, providing automated data collection with daily file organization and pluggable output formats.

Why This Design?

This design allows you to: - Separate data collection logic from file format concerns - Switch output formats (CSV, JSON, Parquet) without changing logging code - Organize data automatically into daily directories with metadata - Handle errors gracefully without stopping the logging process - Reuse the same logger with different storage strategies

Core Components

classDiagram
    %% Core classes
    class SensorLogger {
        -sensor: Sensor
        -tags: List~ReadTag~
        -interval_seconds: int
        -output_path: Path
        -writer: WriterStrategy
        -metadata_saver: SaverStrategy
        -_current_date: date
        -_metadata_written: bool
        -_first_write_today: bool
        +log()
        -_get_daily_dir() Path
        -_write_metadata_if_needed(daily_dir)
        -_validate_interval()
    }

    class Sensor {
        +read(tags) dict~ReadTag, Any~
        +get_valid_tags() List~ReadTag~
    }

    class ReadTag {
        <<abstract>>
        +metadata: ReadTagMetadata
    }

    class WriterStrategy {
        <<interface>>
        +write(filepath, data_point, is_first)
        +get_extension() str
    }

    class SaverStrategy {
        <<interface>>
        +save_logger_metadata(filepath, date, interval, sensor_type, tags)
        +get_extension() str
    }

    %% Concrete implementations
    class CSVWriter {
        +write(filepath, data_point, is_first)
        +get_extension() str
    }

    class JSONMetadataSaver {
        +save_logger_metadata(filepath, date, interval, sensor_type, tags)
        +get_extension() str
    }

    %% Relationships
    SensorLogger o-- Sensor : reads from
    Sensor --o  ReadTag : exposes
    SensorLogger o-- WriterStrategy : formats data with
    SensorLogger o-- SaverStrategy : stores metadata with
    WriterStrategy <|.. CSVWriter : implements
    SaverStrategy <|.. JSONMetadataSaver : implements

How It Works

Logging Flow

When you call logger.log():

1. Daily Directory Management
2. Creates or accesses directory for current date (YYYYMMDD format)

3. Checks if this is the first log of a new day

4. Metadata Handling

5. If first log of the day, writes metadata file using SaverStrategy

6. Metadata includes: date, interval, sensor type, tag information

7. Data Collection

8. Reads all configured tags from the sensor
9. Adds timestamp to create a data point

10. Formats: {timestamp: ISO8601, TagName1: value1, TagName2: value2, ...}

11. Data Writing

12. Writes data point using WriterStrategy

13. Strategy handles file creation, headers, and appending

14. Error Handling

15. Any exceptions are caught and logged to errors.log
16. Logging continues on next call despite errors

Strategy Pattern

The SensorLogger uses the Strategy pattern for flexible I/O operations:

Writer Strategy

Controls how data is written:

from atmospyre.loggers.strategies.writers import CSVWriter

class CustomParquetWriter:
    def write(self, filepath: Path, data_point: dict, is_first: bool):
        # Custom implementation for Parquet format
        ...

    def get_extension(self) -> str:
        return 'parquet'

# Use custom writer
logger = SensorLogger(
    sensor=sensor,
    tags=[CO2, TEMPERATURE],
    interval_seconds=60,
    output_path='./data',
    writer=CustomParquetWriter()  # Plug in your strategy
)

Saver Strategy

Controls how metadata is stored:

from atmospyre.loggers.strategies.savers import JSONMetadataSaver

class YAMLMetadataSaver:
    def save_logger_metadata(self, filepath, date, interval_seconds, sensor_type, tags):
        # Custom implementation for YAML format
        ...

    def get_extension(self) -> str:
        return 'yaml'

# Use custom saver
logger = SensorLogger(
    sensor=sensor,
    tags=[CO2, TEMPERATURE],
    interval_seconds=60,
    output_path='./data',
    metadata_saver=YAMLMetadataSaver()
)

Default Strategies

If no strategy is provided, defaults are used: - Writer: CSVWriter - comma-separated values with headers - Saver: JSONMetadataSaver - JSON formatted metadata