Fix: RainViewer API now uses hash-based paths instead of timestamps

[Goldingr626]

Hi all,
My radar wasn't showing any data just a blank map. I used Claude AI to help me fix the issues.
I also had to make sure Zoom levels in the config file were 7 or smaller.
Hope this helps some people out.

Problem
RainViewer recently changed their API so that radar tile URLs no longer use timestamps as path identifiers. They now use hash-based paths. This causes all radar tiles to return HTTP 410 Gone, resulting in the base map displaying but no radar overlay.

Old URL format:
https://tilecache.rainviewer.com/v2/radar/1776633000/256/7/34/53/6/1_1.png
New URL format:
https://tilecache.rainviewer.com/v2/radar/b9ec619a3e33/256/7/34/53/6/1_1.png
The correct path hash is available from the RainViewer API:
https://api.rainviewer.com/public/weather-maps.json
Additional fixes included

VRB wind direction crash — METAR reports with variable (VRB) wind direction return None for wind_dir, causing a recurring AttributeError crash in wxfinished_metar every 30 minutes
Double slash in tile URLs — tile tails were prefixed with / causing malformed URLs like .../1776633000//256/...

Files changed
PyQtPiClock.py
Testing
Tested on Raspberry Pi running PiClock with Mapbox base maps and RainViewer radar overlay, located in Fort Pierce, Florida (KFPR).

Sure! Here's a clean step-by-step tutorial for fixing a fresh PiClock install.

---

PiClock Radar Fix Tutorial

Step 1 — Fix the wind direction crash

Open the file:

nano ~/PiClock/Clock/PyQtPiClock.py

Press Ctrl+W and search for:

wind_dir.compass

Find this code:

wd = f.wind_dir.compass()
        if Config.wind_degrees:
            wd = str(f.wind_dir.value) + u'°'

Replace it with:

if f.wind_dir is not None:
            wd = f.wind_dir.compass()
            if Config.wind_degrees:
                wd = str(f.wind_dir.value) + u'°'
        else:
            wd = 'VRB'

---

Step 2 — Fix the double slash in tile URLs

Still in the same file, press Ctrl+W and search for:

/256/%d/%d/%d/%d/%d_%d.png

Find this code:

tail = "/256/%d/%d/%d/%d/%d_%d.png" % (self.zoom, x, y,

Change it to (remove the leading slash):

tail = "256/%d/%d/%d/%d/%d_%d.png" % (self.zoom, x, y,

A few lines below, find:

tail = "/256/%d/%d/%d.png?color=%d" % (self.zoom, x, y,

Change it to:

tail = "256/%d/%d/%d.png?color=%d" % (self.zoom, x, y,

---

Step 3 — Fix the RainViewer API (hash-based paths)

Still in the same file, press Ctrl+W and search for:

def getTiles

Find this code:

    def getTiles(self, t, i=0):
        t = int(t / 600)*600
        self.getTime = t
        self.getIndex = i
        if i == 0:
            self.tileurls = []
            self.tileQimages = []
            for tt in self.tiletails:
                tileurl = "https://tilecache.rainviewer.com/v2/radar/%d/%s" \
                    % (t, tt)
                self.tileurls.append(tileurl)

Replace it with:

    def getTiles(self, t, i=0, path=None):
        t = int(t / 600)*600
        self.getTime = t
        self.getIndex = i
        if i == 0:
            if path is None:
                return
            self.tileurls = []
            self.tileQimages = []
            for tt in self.tiletails:
                tileurl = "https://tilecache.rainviewer.com%s/%s" \
                    % (path, tt)
                self.tileurls.append(tileurl)

---

Step 4 — Add the new fetchPathAndGetTiles method

Just above the def getTiles line, add this entire new method:

    def fetchPathAndGetTiles(self, t):
        import urllib2, json
        try:
            r = urllib2.urlopen("https://api.rainviewer.com/public/weather-maps.json")
            d = json.loads(r.read())
            for frame in d['radar']['past']:
                if frame['time'] == t:
                    self.getTiles(t, 0, frame['path'])
                    return
            past = d['radar']['past']
            if past:
                closest = min(past, key=lambda x: abs(x['time'] - t))
                self.getTiles(closest['time'], 0, closest['path'])
        except Exception as e:
            print "fetchPathAndGetTiles error: " + str(e)

---

Step 5 — Update the get() method to call the new function

Press Ctrl+W and search for:

self.getTiles(tt)

Find this code:

            if not gotit:
                self.getTiles(tt)
                break

Replace it with:

            if not gotit:
                self.fetchPathAndGetTiles(tt)
                break

---

Step 6 — Save and restart

Save with Ctrl+O, then exit with Ctrl+X.

Restart PiClock:

cd ~/PiClock
sh startup.sh

Radar should appear within 2-3 minutes as it builds up frames. Done!

[SerBrynden]

We haven’t heard from the original author of this repository, n0bel, in a couple years and we can’t commit changes to it without him.

If your PiClock is not working, I recommend switching to the forked repository at https://github.com/SerBrynden/PiClock. It is up-to-date and working.

There have been some major changes lately with RainViewer weather radar zoom restrictions, color scheme restrictions, and hash-based path format changes for radar API calls.