Garth

Garmin SSO auth + Connect Python client

Google Colabs

Stress: 28-day rolling average

Stress levels from one day to another can vary by extremes, but there's always a general trend. Using a scatter plot with a rolling average shows both the individual days and the trend. The Colab retrieves up to three years of daily data. If there's less than three years of data, it retrieves whatever is available.

Sleep analysis over 90 days

The Garmin Connect app only shows a maximum of seven days for sleep stages—making it hard to see trends. The Connect API supports retrieving daily sleep quality in 28-day pages, but that doesn't show details. Using SleedData.list() gives us the ability to retrieve an arbitrary number of day with enough detail to product a stacked bar graph of the daily sleep stages.

One specific graph that's useful but not available in the Connect app is sleep start and end times over an extended period. This provides context to the sleep hours and stages.

ChatGPT analysis of Garmin stats

ChatGPT's Advanced Data Analysis took can provide incredible insight into the data in a way that's much simpler than using Pandas and Matplotlib.

Start by using the linked Colab to download a CSV of the last three years of your stats, and upload the CSV to ChatGPT.

Here's the outputs of the following prompts:

How do I sleep on different days of the week?

On what days do I exercise the most?

Magic!

Background

Garth is meant for personal use and follows the philosophy that your data is your data. You should be able to download it and analyze it in the way that you'd like. In my case, that means processing with Google Colab, Pandas, Matplotlib, etc.

There are already a few Garmin Connect libraries. Why write another?

Authentication and stability

The most important reasoning is to build a library with authentication that works on Google Colab and doesn't require tools like Cloudscraper. Garth, in comparison:

Uses OAuth1 and OAuth2 token authentication after initial login
OAuth1 token survives for a year
Supports MFA
Auto-refresh of OAuth2 token when expired
Works on Google Colab
Uses Pydantic dataclasses to validate and simplify use of data
Full test coverage

JSON vs HTML

Using garth.connectapi() allows you to make requests to the Connect API and receive JSON vs needing to parse HTML. You can use the same endpoints the mobile app uses.

This also goes back to authentication. Garth manages the necessary Bearer Authentication (along with auto-refresh) necessary to make requests routed to the Connect API.

Instructions

Install

python -m pip install garth

Clone, setup environment and run tests

gh repo clone matin/garth
cd garth
make install
make

Use make help to see all the options.

Authenticate and save session

import garth
from getpass import getpass

email = input("Enter email address: ")
password = getpass("Enter password: ")
# If there's MFA, you'll be prompted during the login
garth.login(email, password)

garth.save("~/.garth")

Custom MFA handler

There's already a default MFA handler that prompts for the code in the terminal. You can provide your own handler. The handler should return the MFA code through your custom prompt.

garth.login(email, password, prompt_mfa=lambda: input("Enter MFA code: "))

Configure

Set domain for China

garth.configure(domain="garmin.cn")

Proxy through Charles

garth.configure(proxies={"https": "http://localhost:8888"}, ssl_verify=False)

Attempt to resume session

import garth
from garth.exc import GarthException

garth.resume("~/.garth")
try:
    garth.client.username
except GarthException:
    # Session is expired. You'll need to log in again

Connect API

Daily details

sleep = garth.connectapi(
    f"/wellness-service/wellness/dailySleepData/{garth.client.username}",
    params={"date": "2023-07-05", "nonSleepBufferMinutes": 60},
)
list(sleep.keys())

[
    "dailySleepDTO",
    "sleepMovement",
    "remSleepData",
    "sleepLevels",
    "sleepRestlessMoments",
    "restlessMomentsCount",
    "wellnessSpO2SleepSummaryDTO",
    "wellnessEpochSPO2DataDTOList",
    "wellnessEpochRespirationDataDTOList",
    "sleepStress"
]

Stats

stress =  garth.connectapi("/usersummary-service/stats/stress/weekly/2023-07-05/52")

{
    "calendarDate": "2023-07-13",
    "values": {
        "highStressDuration": 2880,
        "lowStressDuration": 10140,
        "overallStressLevel": 33,
        "restStressDuration": 30960,
        "mediumStressDuration": 8760
    }
}

Upload

with open("12129115726_ACTIVITY.fit", "rb") as f:
    uploaded = garth.client.upload(f)

Note: Garmin doesn't accept uploads of structured FIT files as outlined in this conversation. FIT files generated from workouts are accepted without issues.

{
    'detailedImportResult': {
        'uploadId': 212157427938,
        'uploadUuid': {
            'uuid': '6e56051d-1dd4-4f2c-b8ba-00a1a7d82eb3'
        },
        'owner': 2591602,
        'fileSize': 5289,
        'processingTime': 36,
        'creationDate': '2023-09-29 01:58:19.113 GMT',
        'ipAddress': None,
        'fileName': '12129115726_ACTIVITY.fit',
        'report': None,
        'successes': [],
        'failures': []
    }
}

Stats resources

Stress

Daily stress levels

DailyStress.list("2023-07-23", 2)

[
    DailyStress(
        calendar_date=datetime.date(2023, 7, 22),
        overall_stress_level=31,
        rest_stress_duration=31980,
        low_stress_duration=23820,
        medium_stress_duration=7440,
        high_stress_duration=1500
    ),
    DailyStress(
        calendar_date=datetime.date(2023, 7, 23),
        overall_stress_level=26,
        rest_stress_duration=38220,
        low_stress_duration=22500,
        medium_stress_duration=2520,
        high_stress_duration=300
    )
]

Weekly stress levels

WeeklyStress.list("2023-07-23", 2)

[
    WeeklyStress(calendar_date=datetime.date(2023, 7, 10), value=33),
    WeeklyStress(calendar_date=datetime.date(2023, 7, 17), value=32)
]

Steps

Daily steps

garth.DailySteps.list(period=2)

[
    DailySteps(
        calendar_date=datetime.date(2023, 7, 28),
        total_steps=6510,
        total_distance=5552,
        step_goal=8090
    ),
    DailySteps(
        calendar_date=datetime.date(2023, 7, 29),
        total_steps=7218,
        total_distance=6002,
        step_goal=7940
    )
]

Weekly steps

garth.WeeklySteps.list(period=2)

[
    WeeklySteps(
        calendar_date=datetime.date(2023, 7, 16),
        total_steps=42339,
        average_steps=6048.428571428572,
        average_distance=5039.285714285715,
        total_distance=35275.0,
        wellness_data_days_count=7
    ),
    WeeklySteps(
        calendar_date=datetime.date(2023, 7, 23),
        total_steps=56420,
        average_steps=8060.0,
        average_distance=7198.142857142857,
        total_distance=50387.0,
        wellness_data_days_count=7
    )
]

Intensity Minutes

Daily intensity minutes

garth.DailyIntensityMinutes.list(period=2)

[
    DailyIntensityMinutes(
        calendar_date=datetime.date(2023, 7, 28),
        weekly_goal=150,
        moderate_value=0,
        vigorous_value=0
    ),
    DailyIntensityMinutes(
        calendar_date=datetime.date(2023, 7, 29),
        weekly_goal=150,
        moderate_value=0,
        vigorous_value=0
    )
]

Weekly intensity minutes

garth.WeeklyIntensityMinutes.list(period=2)

[
    WeeklyIntensityMinutes(
        calendar_date=datetime.date(2023, 7, 17),
        weekly_goal=150,
        moderate_value=103,
        vigorous_value=9
    ),
    WeeklyIntensityMinutes(
        calendar_date=datetime.date(2023, 7, 24),
        weekly_goal=150,
        moderate_value=101,
        vigorous_value=105
    )
]

HRV

Daily HRV

garth.DailyHRV.list(period=2)

[
    DailyHRV(
        calendar_date=datetime.date(2023, 7, 28),
        weekly_avg=39,
        last_night_avg=36,
        last_night_5_min_high=52,
        baseline=HRVBaseline(
            low_upper=36,
            balanced_low=39,
            balanced_upper=51,
            marker_value=0.25
        ),
        status='BALANCED',
        feedback_phrase='HRV_BALANCED_2',
        create_time_stamp=datetime.datetime(2023, 7, 28, 12, 40, 16, 785000)
    ),
    DailyHRV(
        calendar_date=datetime.date(2023, 7, 29),
        weekly_avg=40,
        last_night_avg=41,
        last_night_5_min_high=76,
        baseline=HRVBaseline(
            low_upper=36,
            balanced_low=39,
            balanced_upper=51,
            marker_value=0.2916565
        ),
        status='BALANCED',
        feedback_phrase='HRV_BALANCED_8',
        create_time_stamp=datetime.datetime(2023, 7, 29, 13, 45, 23, 479000)
    )
]

Detailed HRV data

garth.HRVData.get("2023-07-20")

HRVData(
    user_profile_pk=2591602,
    hrv_summary=HRVSummary(
        calendar_date=datetime.date(2023, 7, 20),
        weekly_avg=39,
        last_night_avg=42,
        last_night_5_min_high=66,
        baseline=Baseline(
            low_upper=36,
            balanced_low=39,
            balanced_upper=52,
            marker_value=0.25
        ),
        status='BALANCED',
        feedback_phrase='HRV_BALANCED_7',
        create_time_stamp=datetime.datetime(2023, 7, 20, 12, 14, 11, 898000)
    ),
    hrv_readings=[
        HRVReading(
            hrv_value=54,
            reading_time_gmt=datetime.datetime(2023, 7, 20, 5, 29, 48),
            reading_time_local=datetime.datetime(2023, 7, 19, 23, 29, 48)
        ),
        HRVReading(
            hrv_value=56,
            reading_time_gmt=datetime.datetime(2023, 7, 20, 5, 34, 48),
            reading_time_local=datetime.datetime(2023, 7, 19, 23, 34, 48)
        ),
        # ... truncated for brevity
        HRVReading(
            hrv_value=38,
            reading_time_gmt=datetime.datetime(2023, 7, 20, 12, 9, 48),
            reading_time_local=datetime.datetime(2023, 7, 20, 6, 9, 48)
        )
    ],
    start_timestamp_gmt=datetime.datetime(2023, 7, 20, 5, 25),
    end_timestamp_gmt=datetime.datetime(2023, 7, 20, 12, 9, 48),
    start_timestamp_local=datetime.datetime(2023, 7, 19, 23, 25),
    end_timestamp_local=datetime.datetime(2023, 7, 20, 6, 9, 48),
    sleep_start_timestamp_gmt=datetime.datetime(2023, 7, 20, 5, 25),
    sleep_end_timestamp_gmt=datetime.datetime(2023, 7, 20, 12, 11),
    sleep_start_timestamp_local=datetime.datetime(2023, 7, 19, 23, 25),
    sleep_end_timestamp_local=datetime.datetime(2023, 7, 20, 6, 11)
)

Sleep

Daily sleep quality

garth.DailySleep.list("2023-07-23", 2)

[
    DailySleep(calendar_date=datetime.date(2023, 7, 22), value=69),
    DailySleep(calendar_date=datetime.date(2023, 7, 23), value=73)
]

Detailed sleep data

garth.SleepData.get("2023-07-20")

SleepData(
    daily_sleep_dto=DailySleepDTO(
        id=1689830700000,
        user_profile_pk=2591602,
        calendar_date=datetime.date(2023, 7, 20),
        sleep_time_seconds=23700,
        nap_time_seconds=0,
        sleep_window_confirmed=True,
        sleep_window_confirmation_type='enhanced_confirmed_final',
        sleep_start_timestamp_gmt=datetime.datetime(2023, 7, 20, 5, 25, tzinfo=TzInfo(UTC)),
        sleep_end_timestamp_gmt=datetime.datetime(2023, 7, 20, 12, 11, tzinfo=TzInfo(UTC)),
        sleep_start_timestamp_local=datetime.datetime(2023, 7, 19, 23, 25, tzinfo=TzInfo(UTC)),
        sleep_end_timestamp_local=datetime.datetime(2023, 7, 20, 6, 11, tzinfo=TzInfo(UTC)),
        unmeasurable_sleep_seconds=0,
        deep_sleep_seconds=9660,
        light_sleep_seconds=12600,
        rem_sleep_seconds=1440,
        awake_sleep_seconds=660,
        device_rem_capable=True,
        retro=False,
        sleep_from_device=True,
        sleep_version=2,
        awake_count=1,
        sleep_scores=SleepScores(
            total_duration=Score(
                qualifier_key='FAIR',
                optimal_start=28800.0,
                optimal_end=28800.0,
                value=None,
                ideal_start_in_seconds=None,
                deal_end_in_seconds=None
            ),
            stress=Score(
                qualifier_key='FAIR',
                optimal_start=0.0,
                optimal_end=15.0,
                value=None,
                ideal_start_in_seconds=None,
                ideal_end_in_seconds=None
            ),
            awake_count=Score(
                qualifier_key='GOOD',
                optimal_start=0.0,
                optimal_end=1.0,
                value=None,
                ideal_start_in_seconds=None,
                ideal_end_in_seconds=None
            ),
            overall=Score(
                qualifier_key='FAIR',
                optimal_start=None,
                optimal_end=None,
                value=68,
                ideal_start_in_seconds=None,
                ideal_end_in_seconds=None
            ),
            rem_percentage=Score(
                qualifier_key='POOR',
                optimal_start=21.0,
                optimal_end=31.0,
                value=6,
                ideal_start_in_seconds=4977.0,
                ideal_end_in_seconds=7347.0
            ),
            restlessness=Score(
                qualifier_key='EXCELLENT',
                optimal_start=0.0,
                optimal_end=5.0,
                value=None,
                ideal_start_in_seconds=None,
                ideal_end_in_seconds=None
            ),
            light_percentage=Score(
                qualifier_key='EXCELLENT',
                optimal_start=30.0,
                optimal_end=64.0,
                value=53,
                ideal_start_in_seconds=7110.0,
                ideal_end_in_seconds=15168.0
            ),
            deep_percentage=Score(
                qualifier_key='EXCELLENT',
                optimal_start=16.0,
                optimal_end=33.0,
                value=41,
                ideal_start_in_seconds=3792.0,
                ideal_end_in_seconds=7821.0
            )
        ),
        auto_sleep_start_timestamp_gmt=None,
        auto_sleep_end_timestamp_gmt=None,
        sleep_quality_type_pk=None,
        sleep_result_type_pk=None,
        average_sp_o2_value=92.0,
        lowest_sp_o2_value=87,
        highest_sp_o2_value=100,
        average_sp_o2_hr_sleep=53.0,
        average_respiration_value=14.0,
        lowest_respiration_value=12.0,
        highest_respiration_value=16.0,
        avg_sleep_stress=17.0,
        age_group='ADULT',
        sleep_score_feedback='NEGATIVE_NOT_ENOUGH_REM',
        sleep_score_insight='NONE'
    ),
    sleep_movement=[
        SleepMovement(
            start_gmt=datetime.datetime(2023, 7, 20, 4, 25),
            end_gmt=datetime.datetime(2023, 7, 20, 4, 26),
            activity_level=5.688743692980419
        ),
        SleepMovement(
            start_gmt=datetime.datetime(2023, 7, 20, 4, 26),
            end_gmt=datetime.datetime(2023, 7, 20, 4, 27),
            activity_level=5.318763075304898
        ),
        # ... truncated for brevity
        SleepMovement(
            start_gmt=datetime.datetime(2023, 7, 20, 13, 10),
            end_gmt=datetime.datetime(2023, 7, 20, 13, 11),
            activity_level=7.088729101943337
        )
    ]
)

List sleep data over several nights.

garth.SleepData.list("2023-07-20", 30)

User

UserProfile

garth.UserProfile.get()

UserProfile(
    id=3154645,
    profile_id=2591602,
    garmin_guid="0690cc1d-d23d-4412-b027-80fd4ed1c0f6",
    display_name="mtamizi",
    full_name="Matin Tamizi",
    user_name="mtamizi",
    profile_image_uuid="73240e81-6e4d-43fc-8af8-c8f6c51b3b8f",
    profile_image_url_large=(
        "https://s3.amazonaws.com/garmin-connect-prod/profile_images/"
        "73240e81-6e4d-43fc-8af8-c8f6c51b3b8f-2591602.png"
    ),
    profile_image_url_medium=(
        "https://s3.amazonaws.com/garmin-connect-prod/profile_images/"
        "685a19e9-a7be-4a11-9bf9-faca0c5d1f1a-2591602.png"
    ),
    profile_image_url_small=(
        "https://s3.amazonaws.com/garmin-connect-prod/profile_images/"
        "6302f021-0ec7-4dc9-b0c3-d5a19bc5a08c-2591602.png"
    ),
    location="Ciudad de México, CDMX",
    facebook_url=None,
    twitter_url=None,
    personal_website=None,
    motivation=None,
    bio=None,
    primary_activity=None,
    favorite_activity_types=[],
    running_training_speed=0.0,
    cycling_training_speed=0.0,
    favorite_cycling_activity_types=[],
    cycling_classification=None,
    cycling_max_avg_power=0.0,
    swimming_training_speed=0.0,
    profile_visibility="private",
    activity_start_visibility="private",
    activity_map_visibility="public",
    course_visibility="public",
    activity_heart_rate_visibility="public",
    activity_power_visibility="public",
    badge_visibility="private",
    show_age=False,
    show_weight=False,
    show_height=False,
    show_weight_class=False,
    show_age_range=False,
    show_gender=False,
    show_activity_class=False,
    show_vo_2_max=False,
    show_personal_records=False,
    show_last_12_months=False,
    show_lifetime_totals=False,
    show_upcoming_events=False,
    show_recent_favorites=False,
    show_recent_device=False,
    show_recent_gear=False,
    show_badges=True,
    other_activity=None,
    other_primary_activity=None,
    other_motivation=None,
    user_roles=[
        "SCOPE_ATP_READ",
        "SCOPE_ATP_WRITE",
        "SCOPE_COMMUNITY_COURSE_READ",
        "SCOPE_COMMUNITY_COURSE_WRITE",
        "SCOPE_CONNECT_READ",
        "SCOPE_CONNECT_WRITE",
        "SCOPE_DT_CLIENT_ANALYTICS_WRITE",
        "SCOPE_GARMINPAY_READ",
        "SCOPE_GARMINPAY_WRITE",
        "SCOPE_GCOFFER_READ",
        "SCOPE_GCOFFER_WRITE",
        "SCOPE_GHS_SAMD",
        "SCOPE_GHS_UPLOAD",
        "SCOPE_GOLF_API_READ",
        "SCOPE_GOLF_API_WRITE",
        "SCOPE_INSIGHTS_READ",
        "SCOPE_INSIGHTS_WRITE",
        "SCOPE_PRODUCT_SEARCH_READ",
        "ROLE_CONNECTUSER",
        "ROLE_FITNESS_USER",
        "ROLE_WELLNESS_USER",
        "ROLE_OUTDOOR_USER",
        "ROLE_CONNECT_2_USER",
        "ROLE_TACX_APP_USER",
    ],
    name_approved=True,
    user_profile_full_name="Matin Tamizi",
    make_golf_scorecards_private=True,
    allow_golf_live_scoring=False,
    allow_golf_scoring_by_connections=True,
    user_level=3,
    user_point=118,
    level_update_date="2020-12-12T15:20:38.0",
    level_is_viewed=False,
    level_point_threshold=140,
    user_point_offset=0,
    user_pro=False,
)

UserSettings

garth.UserSettings.get()

UserSettings(
    id=2591602,
    user_data=UserData(
        gender="MALE",
        weight=83000.0,
        height=182.0,
        time_format="time_twenty_four_hr",
        birth_date=datetime.date(1984, 10, 17),
        measurement_system="metric",
        activity_level=None,
        handedness="RIGHT",
        power_format=PowerFormat(
            format_id=30,
            format_key="watt",
            min_fraction=0,
            max_fraction=0,
            grouping_used=True,
            display_format=None,
        ),
        heart_rate_format=PowerFormat(
            format_id=21,
            format_key="bpm",
            min_fraction=0,
            max_fraction=0,
            grouping_used=False,
            display_format=None,
        ),
        first_day_of_week=FirstDayOfWeek(
            day_id=2,
            day_name="sunday",
            sort_order=2,
            is_possible_first_day=True,
        ),
        vo_2_max_running=45.0,
        vo_2_max_cycling=None,
        lactate_threshold_speed=0.34722125000000004,
        lactate_threshold_heart_rate=None,
        dive_number=None,
        intensity_minutes_calc_method="AUTO",
        moderate_intensity_minutes_hr_zone=3,
        vigorous_intensity_minutes_hr_zone=4,
        hydration_measurement_unit="milliliter",
        hydration_containers=[],
        hydration_auto_goal_enabled=True,
        firstbeat_max_stress_score=None,
        firstbeat_cycling_lt_timestamp=None,
        firstbeat_running_lt_timestamp=1044719868,
        threshold_heart_rate_auto_detected=True,
        ftp_auto_detected=None,
        training_status_paused_date=None,
        weather_location=None,
        golf_distance_unit="statute_us",
        golf_elevation_unit=None,
        golf_speed_unit=None,
        external_bottom_time=None,
    ),
    user_sleep=UserSleep(
        sleep_time=80400,
        default_sleep_time=False,
        wake_time=24000,
        default_wake_time=False,
    ),
    connect_date=None,
    source_type=None,
)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

Garth

Google Colabs

Stress: 28-day rolling average

Sleep analysis over 90 days

ChatGPT analysis of Garmin stats

Background

Authentication and stability

JSON vs HTML

Instructions

Install

Clone, setup environment and run tests

Authenticate and save session

Custom MFA handler

Configure

Set domain for China

Proxy through Charles

Attempt to resume session

Connect API

Daily details

Stats

Upload

Stats resources

Stress

Steps

Intensity Minutes

HRV

Sleep

User

UserProfile

UserSettings

Files

README.md

Latest commit

History

README.md

File metadata and controls

Garth

Google Colabs

Stress: 28-day rolling average

Sleep analysis over 90 days

ChatGPT analysis of Garmin stats

Background

Authentication and stability

JSON vs HTML

Instructions

Install

Clone, setup environment and run tests

Authenticate and save session

Custom MFA handler

Configure

Set domain for China

Proxy through Charles

Attempt to resume session

Connect API

Daily details

Stats

Upload

Stats resources

Stress

Steps

Intensity Minutes

HRV

Sleep

User

UserProfile

UserSettings