Channel stats API

Fastapi application that works with parquet files containing meteorological data from different sensors:

• Upload a file, there is a sample in resources folder.
• Request available channels, providing a file identifier and optionally, a type of channel.
• Request statistics for certains channels within a date range.

Endpoints

• /channels/{file_id}

  Retrieve available channel identifiers per type.
  
  

  • Allowed channel types are: vel, std, std_dtr, temp, hum, press, dir, sdir.
    
  • Requesting a channel type outside allowed values will raise an error.
  • Duplicated channel types are filtered before processing.
    
    

  Receives: ChannelRequest model with a file identifier and an optional list of channel_type values.
  Returns: Channels model with dictionary of available channels sorted by channel type.

• /stats/{file_id}

  Retrieve stats (mean and standard deviation) for requested file identifier and channel identifiers.
  
  

  • In case no channel is specified, stats for all channels are returned.
  • A date range may be provided, otherwise stats are computed for the whole time series.
  • Should only start_date be specified, end_date is set to now.
  • Providing any nonexistent channel identifier will raise an error
    
    

  Receives: StatsRequest model with a file identifier and optionally, a list of channels and a date range.
  Returns: Dictionary of Stats models including dictionary of dictionaries with stats sorted by channel.

• /upload

  Upload a parquet file to an object storage backend. Files can only be uploaded once and are assigned an identifier based on the file hash.
  
  

  Receives: Parquet file.
  Returns: FileId model containing file_id and stored values. If an identical file has been uploaded before, stored is false, meaning file was not re-uploaded.

Docker image

• Download image:
  

  docker pull javierjsa/statsapi:latest
  

• Run docker compose:
  

  docker-compose -f docker-compose.yaml up
  

• Access documentation:
  

  http://0.0.0.0/docs