Support for magic method %load

[cpaulik]

Just ran into a notebook that loaded some code from an external python file. Executing this in emacs did nothing.

[millejoh]

Could you attach a copy or provide a link the notebook in question? I haven't used %load much but what I remember is that the search path in the notebook is not always what one would expect.
John

On Fri, Jul 8, 2016 at 9:43 AM -0500, "Christoph Paulik" [email protected] wrote:

Just ran into a notebook that loaded some code from an external python file. Executing this in emacs did nothing.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[cpaulik]

https://github.com/dask/dask-tutorial/blob/master/01-Array.ipynb it's the 5th cell or so down where the first solution is loaded.

[millejoh]

Looks like the behavior should be that the cell executing %load gets its contents replaced by whatever is in the file being loaded. Interesting.

On debugging I can see the following gets sent back by the ipython kernel. I guess EIN needs to start looking for payload attributes in execute_reply messages. Unfortunately free time has been short lately, not sure when I will get around to implementing this.

{
  "metadata": {
    "started": "2016-07-12T20:38:37.066452",
    "engine": "214108de-b808-4ab3-ac01-3b839173611a",
    "status": "ok",
    "dependencies_met": true
  },
  "header": {
    "username": "username",
    "date": "2016-07-12T20:38:37.085454",
    "version": "5.0",
    "msg_type": "execute_reply",
    "msg_id": "3b9af6cc-f057-41a8-8de4-f668469fb007",
    "session": "4419cd77-079a-4112-975f-1e710409c78b"
  },
  "msg_type": "execute_reply",
  "msg_id": "3b9af6cc-f057-41a8-8de4-f668469fb007",
  "channel": "shell",
  "parent_header": {
    "username": "username",
    "date": "2016-07-12T20:38:37.062451",
    "version": "5.0",
    "msg_type": "execute_request",
    "msg_id": "5ad749b4-5a90-45bf-9fe1-ffb314c72270",
    "session": "2acb274f-e232-4855-af0b-0575502166cf"
  },
  "buffers": [],
  "content": {
    "execution_count": 26,
    "payload": [
      {
        "source": "set_next_input",
        "text": "# %load solutions\/Array-01.py\nsums = []\nlengths = []\nfor i in range(0, 1000000000, 1000000):\n    chunk = dset[i: i + 1000000]  # pull out numpy array\n    sums.append(chunk.sum())\n    lengths.append(len(chunk))\n\ntotal = sum(sums)\nlength = sum(lengths)\nprint(total \/ length)\n",
        "replace": true
      }
    ],
    "user_expressions": {},
    "status": "ok"
  }
}

[millejoh]

Okay - think this may be working now. And only a 5 month turnaround! Mom would be so proud.