Helpers to write Luftdaten CSVs to your InfluxDB

README.md

⚠️ This has been moved to https://github.com/tomru/cram-luftdaten ⚠️

Write Luftdaten CSV to a InfluxDB

Helpers to write CSV files from https://www.madavi.de/sensor/csvfiles.php to your InfluxDb.

You can use this to fill up your local InfluxDB with previous data or fill data gaps in your influxdb.

How does it work

• get the relevant CSVs from https://www.madavi.de/sensor/csvfiles.php
• delete all lines that you already have in your DB (you must keep the csv header line)
• export settings as environment vars in case they differ from the defaults (INFLUXDB_HOST, INFLUXDB_PORT, INFLUXDB_DATABASE, SENSOR_ID)
• run cat file1.csv [file2.csv ...] | ./to_line_protocol.py | ./to_influx_db.sh

I should advise you to backup the DB first 😛

to_influx_db.sh

#!/usr/bin/env bash
#
# TODO: explain usage

set -e

HOST=${INFLUXDB_HOST:-localhost}
PORT=${INFLUXDB_PORT:-8086}
DATABASE=${INFLUXDB_DATABASE:-feinstaub}

PRECISION=s

SRC_FILE=${1:-/dev/stdin}

curl --include -X POST \
    "http://${HOST}:${PORT}/write?db=${DATABASE}&precision=${PRECISION}" \
    --data-binary @${SRC_FILE}

to_line_protocol.py

#!/usr/bin/env python3

""" Converts CSV exports from https://www.madavi.de/sensor/csvfiles.php to InfluxDB LineProtocol

    Note: timestamps are in seconds, therefore precision "s" needs to be set
    when writing, see https://docs.influxdata.com/influxdb/v1.2/tools/api/#write

Settings:

    Please set
    * SENSOR_ID
    * INFLUXDB_DATABASE
    in your environment

CSV file spec:

    Semiconlos are used as delimiters. First line are the header columns.
    Generally the column name is used as a DB field name, but there are some legacy
    exceptions like "temperature" instead of "Temp" as listed in the following table.

    | DB Field name      | DB field | CSV Column         | CSV Format
    |--------------------|----------|--------------------|--------------------------
    | none               | time     | Time               | 2017/05/19 00:00:11 (UTC)
    | ?                  | ?        | durP1              |
    | ?                  | ?        | ratioP1            |
    | ?                  | ?        | P1                 |
    | ?                  | ?        | durP2              |
    | ?                  | ?        | ratioP2            |
    | ?                  | ?        | P2                 |
    | SDS_P1             | field    | SDS_P1             | 10.36
    | SDS_P2             | field    | SDS_P2             | 9.50
    | PMS_P1             | field    | PMS_P1             |
    | PMS_P2             | field    | PMS_P2             |
    | temperature        | field    | Temp               | 16.00
    | humidity           | field    | Humidity           | 79.00
    | BMP_temperature    | field    | BMP_temperature    |
    | BMP_pressure       | field    | BMP_pressure       |
    | BME280_temperature | field    | BME280_temperature |
    | BME280_humidity    | field    | BME280_humidity    |
    | BME280_pressure    | field    | BME280_pressure    |
    | samples            | field    | Samples            | 828799
    | min_micro          | field    | Min_cycle          | 172
    | max_micro          | field    | Max_cycle          | 25198
    | signal             | field    | Signal             | -91
    | node               | tag      | N/A filename?      | e.g. esp8266-16229960

"""

import os
import sys
import csv
from datetime import datetime, timedelta


def get_timestamp(timestr):
    """Converts CSV time value to a UTC timestamp in seconds"""
    naive_dt = datetime.strptime(timestr, "%Y/%m/%d %H:%M:%S")
    utc = (naive_dt - datetime(1970, 1, 1)) / timedelta(seconds=1)
    return int(utc)


SENSOR_ID = os.environ.get("SENSOR_ID", "16229960")
DATABASE = os.environ.get("INFLUXDB_DATABASE", "sensors")

NODE = "esp8266-" + SENSOR_ID

NAME_MAP = {
    "Humidity": "humidity",
    "Max_cycle": "max_micro",
    "Samples": "samples",
    "Min_cycle": "min_micro",
    "Signal": "signal",
    "Temp": "temperature",
}

READER = csv.DictReader(sys.stdin, delimiter=";")
for row in READER:
    # error out on legacy format until it's clear what that format is
    if row["Time"] == "time":
        raise Exception(
            "Looks like a legacy format not supported yet. Send the file to the author please."
        )

    # catch multiple column headers
    if row["Time"] == "Time":
        continue

    measurements = []
    for header, value in row.items():
        if header == "Time" or not value:
            continue
        measurements.append("{0}={1}".format(NAME_MAP.get(header, header), value))

    values = {
        "database": DATABASE,
        "node": NODE,
        "measurements": ",".join(measurements),
        "time": get_timestamp(row["Time"]),
    }

    print("{database},node={node} {measurements} {time}".format(**values))

@hristo-stoyanov just tried data-esp8266-2326109-2018-06-14.csv from madavi.de and it worked for me. Did you recently download these files?

As I stated in the comment here I seemed to have had the same problem with some old files (I just forgot). Not sure if all the files on madavi.de have been updated at some point to have the new format.

Best if you compare the column names from old files with the version on madavi.de first.

@hristo-stoyanov

• ignoring multiple csv headers now when using the cat file1.csv file2.csv | approach
• throwing an error when the old format is provided

Feel free to try again.

will continue this on https://github.com/tomru/cram-luftdaten