init

get_data/clean-results.py

@@ -0,0 +1,44 @@
+import subprocess
+import json
+
+with open("./results.json", "r") as f:
+    sports = json.load(f)
+
+ret = {}
+for sport in sports:
+    ret[sport] = {}
+    for event in sports[sport]:
+        ret[sport][event] = {}
+        ret[sport][event]["type"] = "team" if sports[sport][event][0]["participant"]["__typename"] == "GameTeam" else "single"
+        ret[sport][event]["participants"] = [] 
+        print(f"{sport} >> {event} ".ljust(50), end="")
+        for i in range(len(sports[sport][event])):
+            ret[sport][event]["participants"].append({})
+            try:
+                ret[sport][event]["participants"][i]["country"] = sports[sport][event][i]["noc"]["name"]
+            except:
+                try:
+                    ret[sport][event]["participants"][i]["country"] = sports[sport][event][i]["participant"]["countryObject"]["name"]
+                except:
+                    ret[sport][event]["participants"][i]["country"] = "None"
+            if ret[sport][event]["type"] == "single":
+                ret[sport][event]["participants"][i]["athlete"] = {} 
+                ret[sport][event]["participants"][i]["athlete"]["meta_url"] = sports[sport][event][i]["participant"]["meta"]["url"]
+                ret[sport][event]["participants"][i]["athlete"]["name"] = f'{sports[sport][event][i]["participant"]["name"]} {sports[sport][event][i]["participant"]["surname"]}'
+                try:
+                    template = sports[sport][event][i]["participant"]["thumbnail"]["urlTemplate"]
+                    img = f'https://img.olympics.com/images/image/private/t_1-1_300/f_auto{template[template.index("}")+1:]}'
+                except:
+                    img = f'https://gstatic.olympics.com/s1/f_auto/static/light/flag/paris-2024/olympic/3x2/{sports[sport][event][i]["participant"]["countryObject"]["triLetterCode"]}.png'
+                ret[sport][event]["participants"][i]["athlete"]["image"] = img
+            else:
+                try:
+                    ret[sport][event]["participants"][i]["img"] = f'https://gstatic.olympics.com/s1/f_auto/static/light/flag/paris-2024/olympic/3x2/{sports[sport][event][i]["noc"]["code"]}.png'
+                except:
+                    ret[sport][event]["participants"][i]["img"] = None
+            print("o", end="")
+        print()
+
+with open("./data.json", "w") as f: 
+    json.dump(ret, f, ensure_ascii=False, indent=4, sort_keys=True)
+

get_data/get-events.py

@@ -0,0 +1,15 @@
+import subprocess
+import json
+
+with open("./sports.json", "r") as f:
+    sports = json.load(f)
+
+ret = {}
+for sport in sports:
+    stdout = subprocess.check_output(f'./get-events.sh {sport}', shell=True)
+    events = json.loads(stdout)
+    ret[sport] = events
+
+with open("./events.json", "w") as f:
+    json.dump(ret, f, indent=4, ensure_ascii=False, sort_keys=True)
+

get_data/get-events.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+args=("$@")
+
+curl -s -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36" https://olympics.com/en/olympic-games/tokyo-2020/results/${args[0]} |\
+sed -E 's->->\n-g' |\
+rg -A 9999999999999999 '<script id="__NEXT_DATA__" type="application/json">' |\
+sed -n 2p |\
+rev |\
+cut -c10- |\
+rev |\
+jq '[.["props"]["pageProps"]["allEvents"][]["slug"]]'

get_data/get-results.py

@@ -0,0 +1,18 @@
+import subprocess
+import json
+
+with open("./events.json", "r") as f:
+    sports = json.load(f)
+
+ret = {}
+for sport in sports:
+    ret[sport] = {}
+    for event in sports[sport]:
+        stdout = subprocess.check_output(f'./get-results.sh {sport} {event}', shell=True)
+        result = json.loads(stdout)
+        ret[sport][event] = result
+        print(f"Done with {sport} >> {event}")
+
+with open("./results.json", "w") as f:
+    json.dump(ret, f, indent=4, ensure_ascii=False, sort_keys=True)
+

get_data/get-results.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+args=("$@")
+
+curl -s -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36" https://olympics.com/en/olympic-games/tokyo-2020/results/${args[0]}/${args[1]} |\
+sed -E 's->->\n-g' |\
+rg -A 9999999999999999 '<script id="__NEXT_DATA__" type="application/json">' |\
+sed -n 2p |\
+rev |\
+cut -c10- |\
+rev |\
+jq '.["props"]["pageProps"]["gameEvent"]["results"]["standing"][0:9]'

get_data/get.sh

@@ -0,0 +1,12 @@
+#!/usr/bin/env bash
+
+args=("$@")
+
+curl -s -A "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36" https://olympics.com/en/olympic-games/tokyo-2020/results/${args[0]} |\
+sed -E 's->->\n-g' |\
+rg -A 9999999999999999 '<script id="__NEXT_DATA__" type="application/json">' |\
+sed -n 2p |\
+rev |\
+cut -c10- |\
+rev |\
+jq '[.["props"]["pageProps"]["allEvents"][]["slug"]]'