A quick overview on the installation process and usage of the docx-test.py script. Other scripts in this repository are older versions using terribly outdated tactics to achieve the same goal. Contact the creator with question about their processes.
The goal of this project is to save the company time and money by automating major parts of the medical records department. The original way to complete these requisitions would take on average 9 minutes per person. Now it takes 20 seconds.
- Make sure Python version >= 3.10.1 is installed locally
- This requires administrator privileges on company Four Peaks/AUUA machines
- Contact Zack or IT support (XLCON) to install Python on another user to keep OSC user non-admin
- Make sure Python was added to PATH env variable
- Clone the repository to the desired path
- Download Visual Studio Code as an administrator
- Request a Service Account key from the administrator
- Ensure the Account key is stored in the root directory of the cloned repository.
- Obtain the
privKey.pemfrom Zack as the program will not work without it, and it is used to decrypt the information input and output - Install the dependencies from the reqs file using
pip install -r requirements.txt
- Open Visual Studio Code
- Open the folder where the repository was cloned
Ctrl + K & Ctrl + O
- Install the Python extension from the Marketplace
- Press the
button in the top right to run the script within the integrated terminal - Follow the steps!
- The following questions will be prompted to the user
- Fill out the prompted information
- The script will do the following
- Fill out the selected word docs queried by the first question
- Create PDF files out of the documents edited using Py2PDF
- Delete created temporary files (.pdf & .docx)
- Merges and saves the files to a single, correctly named ('Request- Last, First Med Recs Req.pdf') files using PyPDF2 -> (different than Py2PDF!)
- Writes the properties of the
dict-> (patient information) to a new .xlsx file with the same name as the PDF using openpyxl
Learn about more of the cool things Speckle's can do and what he has to offer! This section is mainly for convinience clarity and transparency for HIPAA and other privacy reasons, as is all the source code open source!
Awaiting confirmation to use Phaxio, the current application of faxing comes with the use of Selenium for Python and PyAutoGUI. It uses the development browser created by selenium to manipulate (enter credentials and click buttons) webpages. PyAutoGUI is used to select the most recent file added to the "scans" drive hosted by a computer on the network.
As you can see below, faxing this information through Phaxio would be a much cleaner option, but like with most coding; it can be done 1001 ways.
def fax(patient):
faxNo = patient["fNumber"].replace('.', '')
# forms = ', '.join(patient["forms"])
chromedriver_autoinstaller.install()
driver = webdriver.Chrome(service=Service())
driver.get('https://secure.ipfax.net/')
# Login to the fax service
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div[1]/input').send_keys('username')
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div[2]/input').send_keys('password')
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div[4]/input').click()
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div[5]/div/input').click()
# Fill out the fields
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[1]/div[1]/div/input').send_keys('Medical Records')
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[1]/div[2]/div/input').send_keys(f'{patient["ptName"]} -- {patient["forms"]}')
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[2]/div[1]/div[1]/input').send_keys(faxNo)
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[2]/div[1]/div[2]/input').click()
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[2]/div[7]/button').click()
# Choose file with PyAutoGui
sleep(2)
pag.hotkey('ctrl', 'l')
pag.write(f'S:\\')
pag.press('enter')
sleep(2)
pag.press('tab', 4, 0.25)
pag.press('down')
pag.press('up')
pag.press('enter')
# Sleep to allow ipfax to process file
sleep(2)
# Send the fax
driver.find_element(
by=By.XPATH, value='/html/body/form[1]/div/div/div[2]/div[3]/div[1]/input').click()
sleep(3)
driver.close()Another way to save time creating and sending out medical records is to auto-complete prompts given to the user. These include the surgeon, contacted physician name, phone, and fax numbers.
Starting 1/31/2022 all data (physician name, phone, and fax numbers) submitted by the user will be stored in a Firebase Firestore database to be used by the program further. To provide the ability to use a new function prompt python-prompt-toolkit was the go-to choice for autocompletion.
- The first thing to do is grab all of the data from the database to create the suggestions with the toolkit.
firebase_admin.initialize_app(credentials.Certificate('./sa.json'))
db = firestore.client()
# Create Suggestions
raw_docs = db.collection(u'Auto Suggestions').stream()- Next step is aggregate the data to work with the WordCompleter function.
docs = []
suggestion_list = {'fax': [], 'phone': [], 'dr': [], 'procedure': [], 'surgeons': [
'LaTowsky', 'Mai', 'Kaplan', 'Kundavaram', 'Stern', 'Klauschie', 'Schlaifer', 'Jones', 'Wong', 'Devakumar']}
# Makes the docs from Firestore usable as dicts
for doc in raw_docs:
docs.append(doc.to_dict())
keys = ['fax', 'phone', 'dr', 'procedure']
for doc in docs:
for key in keys:
suggestion_list[key].append(doc[key])- Following the aggregation of the data, implement it in the prompts given to the user.
patient = {
"ptName": prompt(f'What is the name of the patient?\n'),
"dateOfBirth": prompt(f"What is the patient's date of birth?\n"),
"procedureDate": prompt(f'What is the date of the procedure?\n'),
"procedureName": prompt(f'What is the procedure name? (i.e. PVP (Photovaporization of the prostate))\n', completer=FuzzyCompleter(WordCompleter(suggestion_list['procedure'])), complete_in_thread=True, complete_while_typing=True),
"anesthesiologistName": prompt(f'Who is the surgeon? (Kaplan, Wong...)\n', completer=FuzzyCompleter(WordCompleter(suggestion_list['surgeons'])), complete_in_thread=True, complete_while_typing=True),
"drName": prompt(f'What is the name of the doctor you are contacting? (Name only! No "Dr." needed!)\n', completer=FuzzyCompleter(WordCompleter(suggestion_list['dr'])), complete_in_thread=True, complete_while_typing=True),
"pNumber": prompt(f'What is the phone number of the facility you are faxing?\n', completer=FuzzyCompleter(WordCompleter(suggestion_list['phone'])), complete_in_thread=True, complete_while_typing=True),
"fNumber": prompt(f'What is the number you are faxing to?\n', completer=FuzzyCompleter(WordCompleter(suggestion_list['fax'])), complete_in_thread=True, complete_while_typing=True),
"forms": og_prompt["docs"],
"dateOfFax": dt.today().strftime("%B %d, %Y"),
"numberOfPages": str(len(docs)),
"urgency": 'URGENT',
"yourName": 'Names',
}FuzzyCompleteris used to allow the typing of for example 8835 and the suggestions to populate with everything containing 8835... Such as (623 876 8835, 623 883 5824)
- The new suggestions will populate like this
- The 5th and final step is the new suggestions to the database to further improve the suggestions
new_info = {'fax': patient['fNumber'], 'phone': patient['pNumber'],
'dr': patient['drName'], 'procedure': patient['procedureName']}
db.collection('Auto Suggestions').document().set(new_info)Using the Python requests library, it is easy to obtain data from free JSON/REST APIs. There are 4 versions of Speckles...
- Speckles, the Meteorologist 📡
- Provides real-time weather data from OpenWeather API
- Real-feel, cloud cover, rain chance, etc...
- Speckles, the Kanye Enthusiast 🎵
- Gives the user random Ye quotes from Kanye Rest API. Shoutout to the devs of this API, what a gem! @ajzbc on Twitter.
- Speckles, the Activity Planner 🚗
- Explains random activites the user can do if they're bored using Bored API
- Speckles, The Philosopher 🤓
- Provides the user with quotes from books in their data base... QuotePub
Here is the code for giving Speckles his personality!
def speckles():
global responses
weather_speck = json.loads(requests.get(
'https://api.openweathermap.org/data/2.5/weather?q=Sun City&units=imperial&appid=396b8dda92a5079f3bbf2704d32fc382').text)
yeQuotes = []
bookQuotes = []
activities = []
for n in range(5):
# Kayne quotes
quote = json.loads(requests.get('https://api.kanye.rest/').text)
bQuote = json.loads(requests.get(
'https://www.quotepub.com/api/widget/?type=rand&limit=5').text)
yeQuotes.append(quote['quote'])
# Activities
action = json.loads(requests.get(
'https://www.boredapi.com/api/activity').text)
activities.append(
f"With {action['participants']} person/people you can... {action['activity']}.")
bookQuotes.append(
f'"{bQuote[n]["quote_body"]} - {bQuote[n]["quote_author"]}"')
responses = [{
'name': 'Speckles, the Meteorologist 📡',
'prompts': ['<p style="text-align: center; padding: 10px;">🐍 - "Good morning, welcome to the Speckles weather channel!" 😎 </p> ', f'<p style="text-align: center; padding: 10px;">Today is 📅 {dt.today().strftime("%B %d, %Y")} 🐍 - "When is the next holiday again?"</p>', f"<p style='text-align: center; padding: 10px;'>It's feeling like {weather_speck['main']['feels_like']}°F 🐍 - 'Always too damn hot'</p> ", f"The clouds cover {weather_speck['clouds']['all']}% of the sky ☁️ 🐍 - 'Wish it was more...'", f"Get ready for a daily high of: {weather_speck['main']['temp_max']}°F 🤠 'Yee haw'"]
}, {
'name': 'Speckles, the Kanye Enthusiast 🎵 ',
'prompts': yeQuotes
}, {
'name': 'Speckles, the Activity Planner 🚗 ',
'prompts': activities
}, {
'name': 'Speckles, The Philosopher 🤓',
'prompts': bookQuotes
}]After all being created (I may consider doing this asynchronously...) random.choice() chooses one for the rest of the prompts...
def main():
speck_time = random.choice(responses)Then they are added to the prompt's toolbar
"procedureName": prompt(f'What is the procedure name? (i.e. PVP (Photovaporization of the prostate)) 🔪\n', bottom_toolbar=HTML(speck_time["prompts"][0]), completer=FuzzyCompleter(WordCompleter(suggestion_list['procedure'])), complete_in_thread=True, complete_while_typing=True),
"anesthesiologistName": prompt(f'Who is the surgeon? (Kaplan, Wong...) 🩺\n', bottom_toolbar=HTML(speck_time["prompts"][1]), completer=FuzzyCompleter(WordCompleter(suggestion_list['surgeons'])), complete_in_thread=True, complete_while_typing=True),
"drName": prompt(f'What is the name of the doctor you are contacting? (Name only! No "Dr." needed!)\n', bottom_toolbar=HTML(speck_time["prompts"][2]), completer=FuzzyCompleter(WordCompleter(suggestion_list['dr'])), complete_in_thread=True, complete_while_typing=True),
"pNumber": prompt(f'What is the phone number of the facility you are faxing? 📞\n', bottom_toolbar=HTML(speck_time["prompts"][3]), completer=FuzzyCompleter(WordCompleter(suggestion_list['phone'])), complete_in_thread=True, complete_while_typing=True),
"fNumber": prompt(f'What is the number you are faxing to? 📠\n', bottom_toolbar=HTML(speck_time["prompts"][4]), completer=FuzzyCompleter(WordCompleter(suggestion_list['fax'])), complete_in_thread=True, complete_while_typing=True),See below for an in-depth workflow example of the file management used by Speckles
Documentsselected are opened with 'docx' module (Checkbox selection) ->- Following the completion of the
patient,docx_replace_regx()changes the placeholders of the original documents in/medrecs/{filename}.docx-> Documentsare saved as '.docx' files corresponding to their index from the selected documents array ->Documentsare changed to PDFs viadocx2PDFmodule ->- The PDFs are merged together and saved to memory ->
- All .docx and .pdf files are deleted ->
- The merged request PDF is then saved to the current working directory (
cwd)-> file_mgmt.ps1is used to move the documents to the specified directories (backup and '/SCANS') and then delete fromcwd->- A
.xlsxfile is created to be copied into the Clearance Log ✅
Features that are new or in the process of being implemented. These features are currently a Work in Progress, not being used, and still require patching to work properly but are plausible and proof of concepts.
With all of this time saved using this script, the only piece of the puzzle left is to fax and receive files. This can be achieved using an API called Phaxio. The required code to send the file created by this file has been commented out from the main file and can be seen below...
import phaxio
def faxIt(pt):
fileFaxing = f"Request- $Last, $First Med Recs Req.pdf"
phaxio = PhaxioApi('$apiKEY', '$apiSECRET')
phaxio.Fax.send(
to='$receivingNumber',
files=fileFaxing
)
faxIt(patient)