Detecting jams & Notifying People

[matiasandina]

• [x] On FED_library we want to set a timer that lets us know through serial if clearing jam is not solved within 10 min. Send "special alert"
• [X] On FEDWatcher we should be watching for jams. So parse Device_Number, "jam"
• [x] Alert user via email

[matiasandina]

I want to reopen this. We didn't have the best of experience with setting up emails and stuff...there might be a workaround by using something like twillio (or any other app) + python authomation + whatsapp

See here twillio

See here telegram

[matiasandina]

We can send messages with something like this (given we have secrets.py) and have done the telegram configuration. We would need to integrate the proper messaging with parsing data for each FED and/or jamming events and send proper messaging.

Brainstorm of what would be nice to have:

• FED00X is jammed
• FED00X delivered Y number of pellets in the last n hours

import requests
from secrets import secrets
import seaborn as sns
# Send a telegram message to a given user id
def send_message (chatId, message):
    response = requests.post(sendURL + "?chat_id=" + str(chatId) + "&text=" + message)
    # Close to avoid filling up the RAM.
    response.close()

if __name__ == '__main__':
    botToken = secrets['botToken']
    chatId = secrets['telegramDmUid']

    # Telegram send message URL
    sendURL = 'https://api.telegram.org/bot' + botToken + '/sendMessage'

    iris = sns.load_dataset('iris')
    message =  str(iris.head())

    send_message(chatId, message)

[cameron4young]

Here's what I've coded for the runHelper and event_summary functions:

Lines 235-247 (runHelper):

            # TODO: replicate this idea but with the notification interval
            if (now - self.last_notif * 3600) > self.notif_interval:
                self._save_all_df()
                # create read_csv_
                self.last_save = now
                current_directory = os.getcwd() # gets path to current folder (src)
                data_folder_path = os.path.join(current_directory, 'data') # gets the path to the data folder
                csv_files = [file for file in os.listdir(data_folder_path) if file.endswith('.csv')]
                for path in csv_files:
                    summary = self.event_summary(path)
                    if self.tg_enabled:
                        self.send_tg_message(message=summary)
                self.last_notif = now # updates last_notif to the current time

Event summary:

def event_summary(self, path):
    ## want to return message in format "FED{Device #} delivered {num_rows} pellets since {Time}"
    # INPUTS: data, last time interval
    # OUTPUTS: a summarized message in string form

    # iterate through dataframe
    df = pd.read_csv(path)
    device_number = path.split("_")[0]
    # filter through events using PiTime
    filtered_df = df.loc[df['Pi_Time'] > self.last_notif & (df['Pi_Time'] < self.now) & (df['Event']=="Pellet")]
    # pellets = number of rows in df after filtering
    num_pellets = len(filtered_df.index) # will this work with empty df?
    return f"FED{device_number} delivered {num_pellets} since {self.last_notif}"

The function accesses the 'data' folder by using the os.path.join function. A list of paths to all of the csv files are then created. This list is iterated upon, with each file being read and turned into a pandas dataframe within the event_summary function. After the conversion to a pandas dataframe, the dataframe is filtered to only show the data we are interested in, and a summary message is then created and sent. After all of this happens, we then set the last_notif variable to now, and the process repeats every notif_interval.

[matiasandina]

Thanks for the input @cameron4young. As it is, this will not run. Here are a few of parts that I needed to modify. Currently testing.

• now = time.time() (also used for self.last_save) cannot be used to filter a dataframe with a datetime column
• self.now does not exist, created a self.now_dt to make it different from the time.time() version of it which is stored at now path_split will not work as expected when using a full path (which we need for reading!)

These lines are searching for data where the data is not being saved

current_directory = os.getcwd() # gets path to current folder (src)
data_folder_path = os.path.join(current_directory, 'data') # gets the path to the data folder

Data gets saved in

self.exp_dir 

and split by day into

self.today_dir

[matiasandina]

Also, this line turned out to be wrong

filtered_df = df.loc[df['Pi_Time'] > self.last_notif & (df['Pi_Time'] < self.now_dt) & (df['Event']=="Pellet")]
TypeError: Cannot perform 'rand_' with a dtyped [bool] array and scalar of type [bool]

The issue were missing parenthesis on the first condition

---

More importantly this code didn't reset the last_notif and therefore it was blasting notifications one after the other on each iteration. It was missing

self.last_notif = self.now_dt

[matiasandina]

Looks like mvp is working. Closing until further enhancement is needed.