If the user did not provide a log file or folder path, ask: Qual e il percorso del file o della cartella di log messaggi da analizzare?
Prefer passing a folder when the user wants the dashboard updated from all available logs:
```
python {SKILL_DIR}/scripts/generate_message_trend_report.py PATH/TO/LOG_FOLDER -o message_trend.html --json-output message_trend.json
```
The script expands folders recursively to files named log_YYYY_MM_DD_HHMM.txt and generic message logs named log_message.txt, sorted by date and run time after parsing. It ignores derived files such as *_clean.txt and *_timing.txt.

For a single file, run:

python {SKILL_DIR}/scripts/generate_message_trend_report.py PATH/TO/log_YYYY_MM_DD_900.txt -o message_trend.html --json-output message_trend.json

For multiple explicit comparable logs, pass them all in chronological order:

python {SKILL_DIR}/scripts/generate_message_trend_report.py log_2026_04_14_900.txt log_2026_04_15_900.txt -o message_trend.html --json-output message_trend.json

If multiple files exist for the same day, keep them as separate comparable runs. The report labels them as YYYY-MM-DD HHMM, for example 2026-04-21 900 and 2026-04-21 905.
Read message_trend.json after generation and summarize the most useful values in chat:
- analyzed period
- total Packet received
- average per minute
- peak minute and count
- peak second and count
- highest average/minute day or run
- highest peak-minute day or run
- highest peak-second day or run
- top message types
- top machine serials
Confirm the generated files:
- message_trend.html
- message_trend.json

Output Contract

The script generates:

File	Description
`message_trend.html`	Self-contained HTML dashboard for message volume trend
`message_trend.json`	Machine-readable summary used for chat reporting and dashboard data

The JSON includes a highlights object with:

topAvgPerMinute
topPeakMinute
topPeakSecond

Do not move the source log to backup/. Message analysis is non-destructive.

When a new log is added to the folder, rerun the skill on the folder. The dashboard is regenerated from all matching files currently present, so it naturally includes the new file and refreshes same-day comparisons.

Interpreting Results

Use Packet received counts as inbound machine traffic. Packet sent is counted separately, but the dashboard and summary focus on received traffic by default.

Message type is extracted from the MQTT topic segment after CARIMALI/, for example:

Topic prefix	Type
`CARIMALI/DOSEINFO/...`	`DOSEINFO`
`CARIMALI/CONFIG/...`	`CONFIG`
`CARIMALI/VENDING/...`	`VENDING`
`CARIMALI/ALARM/...`	`ALARM`
`CARIMALI/STATUS/...`	`STATUS`

The machine serial is extracted from the next topic segment after the message type.

Manual Fallback

If the script cannot run, inspect lines matching:

[YYYY-MM-DD HH:MM:SS] app.INFO: Packet received [TOPIC]:

Then count:

total received rows
rows grouped by minute (HH:MM)
rows grouped by second (HH:MM:SS)
message types from the topic
serials from the topic

Return the same concise summary in chat. Do not update error_history.json or error_trend.html.

message-log-analizer

Message Log Analizer

Workflow