- Added `normalize_command_content` method to `CommandManager` to streamline command prefix handling, ensuring consistent processing of incoming messages.
- Updated `format_keyword_response`, `check_keywords`, and `execute` methods to utilize the new normalization logic, reducing redundancy and improving clarity.
- Refactored related command classes to remove legacy prefix stripping logic, enhancing maintainability and ensuring all commands adhere to the new normalization approach.
- Enhanced tests to cover various scenarios for command prefix handling, ensuring robust functionality across different configurations.
- Introduced a new class `_SerializedCommands` to serialize host-to-radio commands, ensuring only one command is in-flight at a time to prevent USB-CDC buffer overruns and parser desynchronization.
- Added configuration options for `command_min_interval_ms` and `channel_fetch_interval_ms` to control pacing between commands and channel scans, improving firmware stability during reconnect sequences.
- Updated `ChannelManager` to utilize the new fetch interval during channel scans, and modified the reconnect delay to 10 seconds for gentler recovery from connection issues.
- Enhanced documentation in `config.ini.example` to reflect new settings and their purposes.
NEW_CONTACT adverts were always logged as newly discovered, even for
nodes heard many times before. Check the tracking DB (get_tracked_contact_row)
and the live device contact list (is_contact_on_device) to classify the
advert, log "known" vs "new" accordingly, and skip the
new_contact_discovered audit entry for contacts that are already known.
Also dedup the is_currently_tracked device-contact lookups onto the shared
is_contact_on_device() helper.
Behavior preserved: repeaters remain database-only, device-mode companion
adds are still left to the firmware, and no per-advert device round-trip is
introduced.
- Added new utility functions `espn_dates_for_local_day` and `filter_events_local_day` to handle local timezone date calculations and event filtering.
- Updated `ESPNClient` to support optional date parameters in `fetch_scoreboard_with_calendar` and `fetch_match_states` methods, allowing for more precise data retrieval.
- Modified `WorldCupCommand` and `WorldCupLiveService` to utilize the new date range functionality, ensuring accurate event reporting based on local time.
- Implemented unit tests to verify the correct behavior of date handling and event filtering in various scenarios.
- Added configuration options for announcing red cards, yellow cards, and match stoppages in `config.ini.example`.
- Updated `ESPNClient` to parse and include red and yellow card details in match state data.
- Enhanced `WorldCupLiveService` to announce red and yellow cards, including formatted messages for players and stoppage events.
- Implemented unit tests to validate the new card and stoppage announcement functionalities, ensuring accurate reporting during matches.
- Introduced a new configuration option `announce_disallowed` to allow notifications for goals overturned by VAR.
- Updated `WorldCupLiveService` to handle disallowed goals, including formatting for previously announced scorers.
- Enhanced `ESPNClient` to ensure accurate goal tracking and state management.
- Added unit tests to validate the new disallowed goal functionality and ensure correct behavior during matches.
- Updated ESPNClient to include a detailed list of scoring plays in the live match state, capturing clock, scorer, team ID, own goal, and penalty information.
- Modified WorldCupLiveService to track and announce goals, including formatting for new scorers and handling multiple goals since the last poll.
- Added unit tests to validate goal tracking functionality and ensure accurate announcements during matches.
- Added cache_ttl parameter to fetch_precip_series_nws to enable caching of results for improved performance.
- Implemented logic to reuse cached results based on location and cache expiration.
- Updated tests to ensure proper handling of cache_ttl without causing errors.
- Refactored WorldCupFastcastClient to streamline connection handling and improve readability.
The Open-Meteo forecast model smooths away scattered, pop-up convection, so the rain nowcast (the rain/snow command and the proactive push) could read 0.0 in / ~12% and stay silent while rain was actually falling. Observed near Nashville: Open-Meteo reported 0.0 in across the next 3 h while NWS's own gridpoint showed 65-74% probability with measurable QPF, and thunderstorms were occurring.
Add fetch_precip_series_nws(), which builds the same nowcast-series shape from the NWS gridpoint forecast (6-hour QPF + hourly PoP + weather type). Each hour's precip is its QPF share, zeroed when that hour's PoP is below a floor, so the predicted rain-start tracks the hourly probability instead of snapping to coarse 6-hour QPF boundaries. Both fetchers now prefer NWS for US points and fall back to Open-Meteo where NWS has no coverage (outside the US) or on failure, so the command and the push agree and the model's convective blind spot no longer silences the alert.
Pure helpers (_iso_duration_hours, _nws_hourly, _nws_weather_code) are unit-tested; the NWS-weather -> WMO-code mapping keeps bucket classification (rain/snow/thunder/...) identical to the Open-Meteo path.
- Introduced World Cup command and service settings in `config.ini.example`, `config.ini.minimal-example`, and `config.ini.quickstart` to enable FIFA World Cup score tracking and updates.
- Enhanced `espn_client.py` with methods to fetch tournament calendar and match states, supporting the new World Cup features.
- Updated `sports_command.py` to resolve World Cup nation names and integrate with the new command.
- Added translations for World Cup commands and messages in multiple languages, ensuring comprehensive support for users.
- Included mappings for World Cup nations in `sports_mappings.py` to facilitate accurate team identification.
- Introduced a new static method `_utc_iso_timestamp` to generate UTC ISO 8601 timestamps with a 'Z' suffix for compatibility.
- Updated timestamp generation in `log_packet` and status message to utilize the new method, ensuring consistent timestamp formatting across the service.
- Added a unit test to verify the correct format of the new timestamp method.
Builds on the merged rain/snow nowcast (#193): ten enhancements plus
end-to-end, proactive, and live-smoke test coverage. All new behavior is
config-gated or additive, so existing deployments are unaffected by default.
Enhancements
- Precip amount estimate "(est 0.2 in)" on the command and the proactive push;
snow shown as real depth (Open-Meteo snowfall, cm); freezing rain tagged "in ice".
- Bare country / US state resolves to its capital with a heads-up
(self-contained modules/region_capitals.py; no pycountry/us dependency).
- join_location() dedupes "Spain, Spain" / city-states.
- !snow alias + neutral !nowcast; each looks for its own precip family across
the window, else falls back with a "No snow, but rain ..." cross-type line.
- Keyword-aware help (help rain / help snow).
- Precip probability shown "(..., 70%)"; the proactive incoming alert is gated
to >= [Weather_Service] rain_nowcast_min_probability (default 50).
- Rain<->snow changeover line when the window holds both families.
- Borderline temperature tag (30-38F).
- Short-lived series cache shared by the command and the proactive poll.
Tests
- test_rain_command_e2e.py: drives RainCommand.execute() end to end and asserts
the exact rendered reply across dry/incoming/raining, snow depth, cross-type
mismatch, changeover, ice, temp tag, region capitals, config toggles, DM
budget, and keyword-aware help.
- test_rain_proactive_e2e.py: drives Weather_Service._check_rain_nowcast() for
the probability gate, snow depth, ending notice, and once-per-episode dedup.
- test_rain_live_smoke.py: opt-in (RAIN_LIVE_SMOKE=1) live Open-Meteo check for
upstream schema drift; skipped in CI.
- Shared scaffolding in tests/unit/_rain_harness.py.
New config: [Rain_Command] show_amount/amount_unit/show_probability/show_temp/
cache_seconds/zip_city_lookup; [Weather_Service] rain_nowcast_min_probability/
rain_nowcast_cache_seconds/rain_nowcast_show_amount/rain_nowcast_amount_unit.
ruff + mypy clean.
- Updated the Makefile and GitHub Actions workflow to install ruff version 0.15.15, ensuring consistent linting behavior across environments.
- Added a required-version entry for ruff in pyproject.toml to prevent drift in lint rules.
- Modified the base_service.py method to return None instead of passing, improving clarity.
- Removed an unnecessary blank line in the test_packet_capture_transport_reconnect.py file.
Add a rain nowcast (Open-Meteo 15-minutely precipitation — worldwide, no API
key) as both an on-demand `rain`/`nowcast` command and an opt-in Weather_Service
push that announces rain starting and stopping at the bot's position.
- modules/commands/rain_command.py — command plus pure, unit-tested
fetch/analyze/dedup/label helpers (also reused by the service)
- modules/service_plugins/weather_service.py — background poller mirroring the
existing weather-alert poll pattern; ships disabled (opt-in)
- Location labels resolve to "City, ST" (US) / "City, Country" (non-US)
- 34 unit tests; rain_command added to the strict-mypy module list
- Updated `contacts.html` to escape HTML for contact usernames and advertisement data, preventing potential XSS vulnerabilities.
- Modified `mesh.html` to escape node names and prefixes in tooltips, ensuring safe rendering in the vis-network.
- Removed unnecessary parameters from the `showDeleteConfirmation` method to derive the username directly from the contact data.
- Introduced a new utility function `public_key_has_prefix` to check if a public key starts with a specified prefix in a case-insensitive manner.
- Updated the `PathCommand` class to utilize the new utility function for public key prefix matching, enhancing code readability and maintainability.
- Refactored logic in `PathCommand` and `BotDataViewer` to streamline the handling of recent repeaters based on recency scores, improving clarity in the filtering process.
- Implemented functionality to prevent Discord mention notifications for `@everyone`, `@here`, roles, and users when relaying messages from MeshCore.
- Added a new method to neutralize Discord mention content, ensuring that mentions are displayed as plain text without triggering notifications.
- Updated the Discord webhook payload to include `allowed_mentions` settings to suppress mention parsing.
- Enhanced tests to verify the correct behavior of mention neutralization and webhook payload structure.
- Added `on_transport_reconnected` method to service plugins to re-establish event subscriptions after a transport reconnect.
- Updated `core.py` to notify running services of transport reconnections.
- Enhanced logging for services to track re-subscription actions.
- Added tests to verify the behavior of service plugins during transport reconnects.
- Added configuration options for reconnect behavior, including `reconnect_max_retries`, `reconnect_delay_seconds`, and `reconnect_max_delay_seconds`.
- Enhanced the bot's ability to handle transport disconnects by scheduling reconnect attempts with exponential backoff.
- Updated documentation to reflect new connection settings and behavior.
- Added tests to verify reconnect logic and ensure proper handling of transport errors.
As scopes now become usable in MeshCore, we want to limit the
distribution of the alerts to the regions they are issued for. For that,
we provide a configuration interface to define which region ids
("Regionalschluessel") are mapped to which MeshCore scopes. This reduces
the noise on the warning channels.
- Introduced a new configuration option `multibyte_monitor_enabled` in `config.ini.example` to control the visibility of the multibyte monitor page and API endpoints.
- Implemented the `/multibyte-rollout` and `/api/multibyte-rollout` routes in the web viewer, which are accessible only when the multibyte monitor feature is enabled.
- Updated documentation in `web-viewer.md` to reflect the new feature and its configuration.
- Added tests to ensure the multibyte routes are disabled by default and accessible when enabled.
- Changed the documentation URL in the meshcore-bot.service file to point to the correct repository.
- Updated the project version in pyproject.toml from 0.9.1 to 0.9.2 for consistency in versioning.
- Modified the User-Agent string in dadjoke_command.py to reflect the new repository URL.
- Updated `config.ini.example` and `configuration.md` to enhance clarity on the `outgoing_flood_scope_override` and `flood_scopes` settings, detailing their behavior and interactions.
- Improved logging in `CommandManager` to provide better insights into scope resolution and potential issues during message sending.
- Added new methods in `MessageHandler` for improved RF data correlation, ensuring eligibility checks for flood scope matching.
- Enhanced unit tests to cover new behaviors and ensure robust handling of flood scope configurations.
- Removed global logger level setting and introduced a method to apply log levels based on service-specific verbose/debug settings.
- Added a new method for logging per-packet summaries, allowing for more granular logging control based on verbosity and debug flags.
- Updated packet logging to utilize the new summary method, ensuring appropriate log levels are used for packet capture actions.
Strip leading # when resolving channel names so webhook posts match
hashtag channels cached from the radio. Return HTTP 500 when mesh send
returns false instead of reporting success.
Updated the mesh scope assignment logic in the WebhookService to use a more concise inline conditional expression. This change improves code readability while maintaining the same functionality for determining the appropriate mesh scope based on the incoming request body.
Updated the _send_chunks method to assign timestamps based on the chunk index, ensuring each chunk receives a unique timestamp that increments by one second. This change facilitates proper ordering and deduplication of messages on the client side. Added a unit test to verify the correct behavior of timestamp assignment during chunk transmission.
Updated the GitHub Actions workflow to exclude the dist directory from shellcheck checks, ensuring cleaner linting results.
fix(command_manager): reorder datetime import
Moved the datetime import to the correct position in command_manager.py for better code organization.
fix(reload_config): improve config value retrieval
Refactored the _read_config_value function in reload_config.sh to ensure default values are returned correctly when the key is not found in the config file.
fix(tests): adjust imports in test_scheduler_logic
Updated import statements in test_scheduler_logic.py for consistency and clarity.
- Introduced optional regional TC_FLOOD scope configuration for various services, allowing for more granular control over message routing.
- Updated the CommandManager to resolve and apply flood scopes from messages, configuration sections, and explicit parameters, ensuring correct message delivery.
- Enhanced service plugins to utilize the new flood scope functionality, including weather, earthquake, and webhook services.
- Added unit tests to verify the correct resolution and application of flood scopes in different scenarios, ensuring robust functionality.
Updated the CommandManager to prioritize sender_pubkey over sender_id when sending direct messages. This change ensures correct routing of messages, especially in cases where multiple nodes share similar display names. Added tests to verify the new behavior and ensure proper functionality.
For the emergency communication, we need to ensure that all messages are
transmitted (and received) correctly. We already check if a repeater
acks our message, however a missing ack does not tell us if the message
was lost, or the ack. For that, we repeat the message in this case,
leading to potentially content-wise duplicated receptions on the network.
To let the network depulicate the traffic, the re-transmissions need to
be bit identical with the original message. For that, we compute a
timestamp on initial send and use the same timestamp on
re-transmissions (per message). This further allows clients to
chronologically sort the received messages, making the result easier to
read.
Signed-off-by: Felix Moessbauer <felix.moessbauer@gmail.com>
By that, bit identical replications of channel messages can be created,
which is needed if re-transmissions are performed that should be
deduplicated by the receiver. It further is possible (if the client
supports this) to control the display order of the messages
(chronological).
Signed-off-by: Felix Moessbauer <felix.moessbauer@gmail.com>
Some services are named <foo>_Service, whereby currently only the
Service part is stripped. By that, the service name is derived to be
<foo>_, instead of <foo>.
We fix this by stripping all leading and trailing underscores in the
service name.
Signed-off-by: Felix Moessbauer <felix.moessbauer@gmail.com>
- Introduced `jwt_ttl_seconds` and `jwt_renewal_interval` settings in `config.ini.example` for global JWT management, allowing for better control over token expiration and renewal intervals.
- Updated documentation in `packet-capture.md` to clarify the usage of global and per-broker JWT settings, enhancing user understanding of authentication configurations.
- Refactored `PacketCaptureService` to incorporate new JWT settings, ensuring consistent handling of token lifetimes and renewal processes.
- Updated the `track_contact_advertisement` method to return a structured result indicating success and duplicate packet status, improving clarity in handling contact advertisements.
- Adjusted the `MessageHandler` to utilize the new result structure, allowing for better decision-making based on tracking outcomes.
- Enhanced unit tests to cover new behaviors and ensure correct handling of duplicate packets and advertisement tracking results.
Scheduled cron jobs now pass the cron key into the send path, apply a
deterministic [0, N) second delay from scheduled_message_max_stagger_seconds
(default 1.5), and call send_channel_message with skip_user_rate_limit=True
so simultaneous jobs are not dropped by rate_limit_seconds. Per-channel and
bot_tx limits still apply. Document the new Bot key in config.ini.example;
tests assert kwargs, skip flag, and stagger behavior.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Simplified the flushing mechanism for pending updates in the MeshGraph class by introducing a `flush_needed` flag to reduce redundant calls to `_flush_pending_updates_sync()`.
- Enhanced the FeedManager to prevent coroutine pileup by checking the lock status before acquiring it, allowing for more efficient message processing.
- Introduced a persistent dictionary to track the last send time for each feed, ensuring send intervals are respected without blocking.
- Updated the MessageScheduler to fire-and-forget the message processing, improving responsiveness and logging errors without blocking the main thread.
- Introduced `reply_prefix` to prepend a customizable string to path command replies, supporting placeholders for dynamic content.
- Added `minimum_path_bytes` setting to control the resolution of repeater names based on the number of bytes per hop, enhancing path command behavior.
- Updated relevant documentation and translations to reflect these new configuration options.
- Implemented unit tests to ensure correct functionality of the new features.
Route process_message keyword/RandomLine sends through send_response so
reply_scope reaches set_flood_scope; add optional command_id to
send_response for transmission tracking.
- Updated the scheduled message format in `config.ini.example` to support 5-field cron expressions and preset aliases, replacing the deprecated HHMM format.
- Improved the `MessageScheduler` class to parse and validate new schedule formats, logging warnings for deprecated usage.
- Adjusted the `ScheduleCommand` to display scheduled messages with their respective cron or preset labels.
- Added unit tests to ensure correct parsing and handling of various schedule formats, including legacy HHMM and cron expressions.
- Replaced the use of AdvertFlags enum for flag parsing with bitwise operations to handle invalid values correctly.
- Updated the parsing logic in both MessageHandler and MapUploaderService to use boolean flags for better clarity and performance.
- Adjusted error logging to use warnings instead of errors for parsing issues, enhancing the logging strategy.
- Added unit tests to ensure correct parsing behavior for edge cases with invalid flag values.
Add bridge_outbound posting, BaseServicePlugin.send_external_notifications,
and RepeaterPrefixCollision discovery vs collision routing with silence_mesh_output.
- Updated string quotes from single to double quotes for consistency across the MessageHandler class.
- Enhanced formatting of method parameters for better readability.
- Minor adjustments to logging statements for clarity.
- Added configuration options `mqtt_skip_unparseable_packets` and `advert_require_valid_signature` to control MQTT publishing behavior based on packet validity.
- Updated `decode_path_len_byte` function to return `None` for reserved size codes, improving path length validation.
- Implemented logic in `PacketCaptureService` to skip publishing unparseable packets and ADVERT packets with invalid signatures.
- Introduced `verify_meshcore_advert_ed25519` function for signature verification of ADVERT packets, with corresponding unit tests to ensure functionality.
- Enhanced documentation to reflect new configuration options and their effects on packet processing.
agessaman