As far as I can test, SDA and SCL on Wemos D1 (D1 B, D2 C) are not connected to the I2C Grove and Gravity pins. Is this possible or I'm missing something?

Hi, pin spacing for the Grove connectors is 2mm.
The pin spacing in the board is the standard 2.54mm instead, so a connector doesn't fit unless you bend its pins.

From Programmable WEB:

the call is based on this Documentation

`
"""
The purpose of this code is to show how to work with plant.id API.
You'll find API documentation at https://plantid.docs.apiary.io and https://plant.id/api
"""

import base64
import requests
from time import sleep

secret_access_key = '-- ask for one at [email protected] --'

def send_for_identificattion(file_names):
files_encoded = []
for file_name in file_names:
with open(file_name, 'rb') as file:
files_encoded.append(base64.b64encode(file.read()).decode('ascii'))

params = {
	'latitude': 49.194161,
	'longitude': 16.603017,
	'week': 23,
	'images': files_encoded,
	'key': secret_access_key,
	'parameters': ["crops_fast"]
}
# see the docs for more optinal atrributes; for example 'custom_id' allows you to work
# with your own identifiers
headers = {
	'Content-Type': 'application/json'
}

response = requests.post('https://plant.id/api/identify', json=params, headers=headers)

if response.status_code != 200:
    raise("send_for_identificattion error: {}".format(response.text))

# this reference allows you to gather the identification result (once its ready)
return response.json().get('id')

def get_suggestions(request_id):
params = {
"key": secret_access_key,
"ids": [request_id]
}
headers = {
'Content-Type': 'application/json'
}

# To keep it simple, we are pooling the API waiting for the server to finish the identification.
# The better way would be to utilise "callback_url" parameter in /identify call to tell the our server 
# to call your's server enpoint once the identificatin is done.
while True:
	print("Waiting for suggestions...")
	sleep(5)
	resp = requests.post('https://plant.id/api/check_identifications', json=params, headers=headers).json()
	if resp[0]["suggestions"]:
		return resp[0]["suggestions"]

more photos of the same plant increases the accuracy

request_id = send_for_identificattion(['photo1.jpg','photo2.jpg'])

just listing the suggested plant names here (without the certainty values)

for suggestion in get_suggestions(request_id):
print(suggestion["plant"]["name"])
`