lambda_function.py

import json
import ast
import pandas as pd
import requests
import textwrap
import simplejson
from datetime import datetime
print('Loading function')

LINE_LENGTH = 15
TESTING = False # True

def lambda_handler(event, context):
    jsonp_body = "// This is crusty, but it seems to work.\n// Generated at {}".format(str(datetime.now()))
    ret_dict = {}

    def mergeD(ugly_data):
        """Put the ID inside the dictionary that becomes the row."""
        f = ugly_data["fields"]
        i = {"id": ugly_data["id"]}
        f.update(i)
        return f
    
    if TESTING:
        t = mergeD({'createdTime': '2018-05-25T07:45:33.000Z', 
                    'fields': {'Other_people_involved': ['rece9qyNSfMqAhZAL'],
                               'Owner': ['recgcDkgKvrBLH1k0'], 
                               'LongName': 'A Plane of Thrones Paper', 
                               'Name': 'potPaper', 
                               'percentComplete': 100, 
                               'precursors': ['recflCm8BzuGQbmW0', 'recSIIp1grVggDGQo'], 
                               'Type': 'Publication'}, 
                    'id': 'rec0ArDcPS4hLrmJq'})
        print(t)
    
    
    def get_table(table_name):
        """Get the table from the OATMEAL airtable database and return it as a DataFrame."""
        auth_headers = {'Authorization': 'Bearer keyYXkjYFw61SeWDk'}
        table_url = 'https://api.airtable.com/v0/appuaXpFiadmP89sq/' + table_name
        r = requests.get(table_url, headers=auth_headers)
        d = eval(ast.literal_eval(str(r.content)[1:])) # this is almost certainly not the right way to do this!
        
        rows = [mergeD(x) for x in d["records"]]
        data_frame = pd.DataFrame(rows)
        return data_frame
    
    if TESTING:
        t = get_table("Techs")
        print(t)
    
    
    def pull_out_embed_url(embed_code):
        """Pull the embed url out of the iframe link."""
        if type(embed_code) is not float:
            for thing in embed_code.split(" "):
                if "src=" in thing:
                    thing = thing.split("\"")
                    for part in thing:
                        if "https://" in part:
                            return part.split("?")[0]
        else:
            return "https://www.youtube.com/embed/dQw4w9WgXcQ?rel=0&showinfo=0"
        
    if TESTING:
        embed = ('<iframe '
                 'src="https://docs.google.com/presentation/d/e/'
                 '2PACX-1vRF7DSeGnbLr_zwWnHVrUF0ptxjfIoj9bDwkaJb370IYYWEwF48yszwLUUTAECQ5F6mlJMrUrlQqdj-/embed?'
                 'start=false&loop=false&delayms=60000"'
                 'frameborder="0"'
                 'width="960"'
                 'height="569"'
                 'allowfullscreen="true"'
                 'mozallowfullscreen="true"'
                 'webkitallowfullscreen="true"></iframe>')
        print( pull_out_embed_url(embed) )
    
    
    def wrap_long_name(row):
        """Insert line breaks to make names fit in the node."""
        try:
            ln = "\n".join( textwrap.wrap(row["LongName"], LINE_LENGTH))
        except Exception as e:
            print(e)
            ln = row["LongName"]
        return ln
    
    
    def make_node_description(row, long_name_broken):
        node_description_pattern = "{name} [label=\"{longName}\"];"
        return node_description_pattern.format(name=row["Name"], longName=long_name_broken)
    
    
    project_data = get_table("Techs")
    project_data.set_index("id", inplace=True)
    project_data.deck_embed = project_data.deck_embed.apply(pull_out_embed_url)
    
    group_data = get_table("groups")
    group_dict = {}
    for _, row in group_data.iterrows():
        print(row)
        group_dict[row.id] = {"name":row.Name,
                              "longName": row.LongName,
                              "nodes":[]
                             }

    edge_pattern = "{from_node} -> {to_node};"
    
    node_descriptions = []
    edges = []
    for _, row in project_data.iterrows():
        ln = wrap_long_name(row)
        n = make_node_description(row, ln)
        node_descriptions.append(n)
        
        if type(row["precursors"]) is not float: # nothing reads as NaN, therefore float
            precursors = row["precursors"]#.split(",")
            for precursor in precursors:
                e = edge_pattern.format(from_node=project_data.loc[precursor].Name,
                                        to_node=row["Name"])
                edges.append(e)
        
        if type(row["Group"]) is not float: # nothing reads as NaN, therefore float
            id = row["Group"][0]
            print(group_dict[id])
            group_dict[id]["nodes"].append(row["Name"])
    
    # print(group_dict)
    subgraph_template = """
subgraph cluster_{name} {{
    style="rounded"
    node [style=filled, color=white];
    {nodes};
    label = "{long_name}";
}}"""
    subgraphs = ""
    for _, v in group_dict.items():
        subgraphs += subgraph_template.format(name=v["name"],
                                              long_name=v["longName"],
                                              nodes="; ".join(v["nodes"]))
    
    
    g = """digraph G {{
      size = "16.66,8.33!"; // 1200x600 at 72px/in, "!" to force
      ratio = "fill";
      rankdir  = "LR";
      label    = "Tech Tree of OATMEAL Projects";
      labelloc = "t";
      node [style=filled,
            shape="oval", 
            fixedsize=true, 
            width=1,
            height=1,
            fontsize=8
           ];
      edge [arrowhead=vee,
            arrowsize=0.5]
      
      {nodes}
      
      {edges}

      {subgraphs}
      
      //{{rank = same; data; basicRobot; hardware; systemReef; SNAawareness}};
      
    }}""".format(edges="\n  ".join(edges), 
                 nodes="\n  ".join(node_descriptions),
                 subgraphs=subgraphs
                 ) 
    
    
    # print(g)
    
    
    ret_dict['digraph_js'] = g  # TODO check if new lines are needed
    
    
    simple_dict = project_data.to_dict(orient="records")
    new_dict = {}
    for row in simple_dict:
        new_dict[row["Name"]] = row
    
    j = simplejson.dumps(new_dict, ignore_nan=True, sort_keys=True, indent=2)
    ret_dict["parallelData"] = j    
    jsonp_body += "\n\n\n" + j
        
    
    def get_thumb(row):
        try:
            h = row["Headshot"]
            if type(h) is not float:
                # TODO: make this check for small as well
                x = list(h)
                return x[0]['thumbnails']["large"]["url"]
            else:
                return "http://getdrawings.com/images/casper-drawing-8.jpg"
        except Exception as e:
            print(e)
            return "http://getdrawings.com/images/casper-drawing-8.jpg"
        
    
    people_data = get_table("People")
    people_data["thumb_large"] = people_data.apply(get_thumb, axis=1)
    
    # make the dictionary have the project name as a key, instead of bing a list of independent dicts
    simple_people_dict = people_data.to_dict(orient="records")
    new_p_dict = {}
    for row in simple_people_dict:
        new_p_dict[row["id"]] = row
        
    new_p_dict
    
    j = simplejson.dumps(new_p_dict, ignore_nan=True, sort_keys=True, indent=2)
    ret_dict['peopleData'] = j
    # j = "var peopleData = {json};".format(json=j)
    # with open('peopleData.js', 'w') as f:
    #     f.write(j)
    # jsonp_body += "\n\n\n" + j
    return ret_dict
            #raise Exception('Something went wrong')

if TESTING:
    lambda_handler(1,2)