Merge branch 'main' into juan/2

Author: jevillegasd

.pre-commit-config.yaml

@@ -22,7 +22,7 @@ repos:
           - --config=pyproject.toml
           - --all
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.12.0
+    rev: v0.12.5
     hooks:
       # Run the linter.
       - id: ruff

doc/source/protocols/readout_optimization.rst

@@ -0,0 +1,102 @@
+Readout optimization
+====================
+
+Qibocal provides different protocols to improve the readout pulse parameters by optimizing the assignment fidelity.
+
+Resonator Amplitude
+-------------------
+
+The resonator amplitude protocol improves the readout amplitude by optimizing the assignment fidelity.
+
+Parameters
+^^^^^^^^^^
+
+.. autoclass:: qibocal.protocols.readout_optimization.resonator_amplitude.ResonatorAmplitudeParameters
+	:noindex:
+
+
+Example
+^^^^^^^
+
+It follows an example runcard of the resonator amplitude routine with the plot
+generated in the report.
+
+.. code-block:: yaml
+
+    - id: resonator_amplitude
+      operation: resonator_amplitude
+      parameters:
+          amplitude_step: 0.0005
+          amplitude_start: 0.001
+          amplitude_stop: 0.005
+
+As shown in the picture below, the protocol sweeps the readout amplitude and evaluates the errors probability.
+
+.. image:: readout_amplitude.png
+
+
+Resonator Frequency
+-------------------
+
+The resonator frequency protocol performs a sweep of the readout frequency to maximize assignment fidelity.
+
+
+Parameters
+^^^^^^^^^^
+
+.. autoclass:: qibocal.protocols.readout_optimization.resonator_frequency.ResonatorFrequencyParameters
+	:noindex:
+
+
+Example
+^^^^^^^
+
+In the following we show a possible runcard for the optimization of the resonator frequency together with expected output plot.
+
+.. code-block:: yaml
+
+    - id: resonator_frequency
+      operation: resonator_frequency
+      parameters:
+        freq_width: 1_000_000
+        freq_step: 100_000
+
+.. image:: frequency_optimization.png
+
+
+
+Resonator Optimization
+----------------------
+
+In addition to this metric, the protocol also computes the quantum non-demolition-ness (QND) of the measurement, which serves to evaluate measurement-induced state disturbance.
+To capture leakage outside the computational basis the protocol includes the computation of QND with an additional :math:`\pi` pulse.
+Readout parameters are then updated based on the configuration that yields the highest assignment fidelity.
+
+
+Parameters
+^^^^^^^^^^
+
+.. autoclass:: qibocal.protocols.readout_optimization.resonator_optimization.ResonatorOptimizationParameters
+	:noindex:
+
+
+Example
+^^^^^^^
+
+A possible runcard for the resonator optimization protocol could be:
+
+.. code-block:: yaml
+
+    - id: resonator optimization
+      operation: resonator_optimization
+      parameters:
+        freq_width: 3_000_000
+        freq_step: 125_000
+        amplitude_step: 0.00025
+        amplitude_start: 0.00025
+        amplitude_stop: 0.0085
+        delay: 1000
+
+The plot generated by the report should be similar to the following one:
+
+.. image:: resonator_optimization.png

doc/source/protocols/readout_optimization/frequency_optimization.png

@@ -9,7 +9,15 @@ print(s.getsockname()[1])
 s.close()
 EOF
 
-IP=$(hostname -I | cut -d ' ' -f 1)
+read -rd '' getip <<EOF
+import sys
+
+ips = sys.stdin.read().split(" ")
+print([ip for ip in ips if ip.startswith(sys.argv[1])][0])
+EOF
+
+SUBNET=""
+IP=$(hostname -I | python -c "$getip" "$SUBNET")
 PORT=$(python -c "$getport")
 
 echo "Forward your local port with:"
@@ -19,6 +27,6 @@ echo
 echo "Connect to:"
 echo "  https://localhost:9180"
 
-jupyter lab --ip "$IP" --port "$PORT"
+jupyter lab --ip "$IP" --port "$PORT" --no-browser
 # you can also test this with:
 # python -m http.server -b "$IP" "$PORT"

…c/source/protocols/readout_amplitude.png …adout_optimization/readout_amplitude.pngdoc/source/protocols/readout_amplitude.png renamed to doc/source/protocols/readout_optimization/readout_amplitude.png doc/source/protocols/readout_amplitude.png renamed to doc/source/protocols/readout_optimization/readout_amplitude.png

@@ -33,3 +33,60 @@ the following script should help automating the process.
 
 Once the required server is dispatched as queue job with the script, the instructions to
 connect will appear in the output (usually redirected by the queue system on a file).
+
+
+.. note::
+
+   If the system you are targeting, the *remote computing node*, is actually part of
+   multiple networks, and you need to discriminate the correct one to use, just add the
+   subnetwork as a string in the former
+
+   E.g. you could use::
+
+       SUBNET="192.168.XXX"
+
+   to specify a local network whose 3rd byte is ``XXX`` (e.g. ``0``, ``1``, ..., ``255``)
+
+
+Step-by-step guide
+~~~~~~~~~~~~~~~~~~
+
+In case of doubts, let's go step by step through the script usage.
+
+#. the **script** itself is supposed to be executed on the **remote computing node**
+   (cf. figure above), which is the one with direct access to the resources (thus, the
+   one which is usually already behind the queue)
+#. notice that the instructions printed in the output will direct you to establish a
+   **further SSH connection**, which will be the bridge between the local system and the
+   remote computing node, through the access node (cf. figure above), on it can be
+   established with a further `ssh` command execution on the **local system**
+#. eventually, the **client connection** should be established from the same **local
+   system**
+
+.. admonition:: Example: SLURM + laptop
+
+   In the case of SLURM queue system, which is accessible through an access node, where
+   the user `ssh` from its laptop:
+
+   - the script is the one you should `sbatch` - possibly adding to it all the rest of
+     the logic, as needed
+   - the further `ssh` command (printed in the instructions) should be run on the laptop
+   - the final URL to connect can be pasted in a browser on the same laptop (e.g.
+     Mozilla Firefox)
+
+Further notes
+~~~~~~~~~~~~~
+
+.. admonition:: Jupyter tokens
+
+   *This is not related to this guide, but just to Jupyter itself.*
+
+   When Jupyter opens the browser automatically, it uses a URL containing a *token*, for
+   ease of access.
+   If you just specify the address and port, the token will not be included. But you can
+   find it in the output of the command in the jupyter server (thus in the log of the
+   queue job, if you are using a queue, as in this guide), and manually insert it in the
+   page which is presented by Jupyter when the URL token is not used.
+
+   More details in Jupyter docs:
+   https://jupyter-server.readthedocs.io/en/stable/operators/security.html

doc/source/protocols/readout_optimization/readout_optimization.rst

@@ -25,6 +25,11 @@
 
           env = {
             QIBOLAB_PLATFORMS = (dirOf config.env.DEVENV_ROOT) + "/qibolab_platforms_qrc";
+            LD_LIBRARY_PATH = builtins.concatStringsSep ":" (map (p: "${p}/lib") (with pkgs; [
+              stdenv.cc.cc.lib
+              zlib
+            ]));
+            PYTHONBREAKPOINT = "pudb.set_trace";
           };
 
           languages = {

doc/source/protocols/readout_optimization/resonator_optimization.png

@@ -12,7 +12,7 @@
   },
   "hamiltonian":{
     "transmon_levels": 2,
-    "single_qubit": {
+    "qubits": {
       "0": {
         "frequency": 5e9,
         "anharmonicity": -200e6,
@@ -74,7 +74,7 @@
             "0/drive",
             {
               "duration": 40,
-              "amplitude": 0.07975,
+              "amplitude": 0.06349,
               "envelope": {
                 "kind": "gaussian",
                 "rel_sigma": 0.2
@@ -97,9 +97,8 @@
                 "duration": 100.0,
                 "amplitude": 0.1,
                 "envelope": {
-                  "kind": "gaussian_square",
-                  "rel_sigma": 0.2,
-                  "width": 0.75
+                  "kind": "gaussian",
+                  "rel_sigma": 0.2
                 },
                 "relative_phase": 0.0,
                 "kind": "pulse"