removed num_atoms + updated docs

docs/source/protocol_proposed.rst

@@ -3,9 +3,8 @@ Proposed Protocol v3
 
 The suggested changes to the protocol are as follows:
 
-1. Via an ``.mdp`` file setting, the user should be able to specify which of positions, forces, and velocities are sent.
-   These do not need to be rate-adjustable like with xtc and trr output settings- they will always be output vai IMD for every frame if
-   turned on.
+1. Via an ``.mdp`` file setting, the user should be able to specify which of simulation box, positions, forces, and velocities are sent.
+   These should be rate-adjustable like with xtc and trr output settings but should be set to every step by default.
 
 2. The IMD_HANDSHAKE packet should have a 1-byte body which contains the configuration settings the simulation was setup with.
    This will allow the client to be choose appropriate packets to send and receive without redundant configuration.
@@ -21,16 +20,20 @@ The suggested changes to the protocol are as follows:
         <val> (bit) (imdpull: true or false)
         <val> (bit) (imdwait: true or false)
         <val> (bit) (imdterm: true or false)
+        <val> (bit) (wrapped positions: true or false)
         <val> (bit) (positions included: true or false)
         <val> (bit) (velocities included: true or false)
         <val> (bit) (forces included: true or false)
         <val> (bit) (dimensions included: true or false)
-        <val> (bit) (unused bit)
+        
+        "wrapped positions" will be a new ``.mdp`` setting which specifies whether the atoms' positions
+        should be adjusted to fit within the simulation box before sending. This is useful for visualization purposes.
 
 3. The server should wait longer than 1 second (possibly up to 60s) for the go signal so that the client 
-   has plenty of time to allocate memory buffers based on the endianness and data type information it received in the handshake packet.
+   has plenty of time to allocate memory buffers based on the endianness and information on included data types 
+   it received in the handshake packet.
 
-4. In the simulation loop, the server will send the client data in this order (iff the configuration says to send it)
+4. In the simulation loop, the server will send the client data in this order (if the configuration says to send it)
 
     i. Dimension data (IMD_DIM) in triclinic vectors
 
@@ -40,14 +43,5 @@ The suggested changes to the protocol are as follows:
 
     iv. Force data (IMD_FORCES) in the same manner as positions
 
-5. The server will no longer send positions adjusted for visualiziation purposes, they will instead be sent in the same manner as .xtc or .trr
-   positions are written to files. The client will be responsible for this adjustment since the dimension data is now available.
-
-6. The server will send a new IMD_EOS (end of stream) packet after the last frame is sent unless the client initiates the disconnection with
+5. The server will send a new IMD_EOS (end of stream) packet after the last frame is sent unless the client initiates the disconnection with
    IMD_DISCONNECT.
-
-
-Questions
-^^^^^^^^^
-
-- Should the server provide the client with the number of atoms?

imdreader/IMDREADER.py

@@ -21,8 +21,7 @@
     import MDAnalysis as mda
     # Pass host and port of the listening GROMACACS simulation
     # server as the trajectory argument
-    # Number of atoms must be provided for internal buffer allocation
-    u = mda.Universe("topology.tpr", "localhost:8888", num_atoms=100)
+    u = mda.Universe("topology.tpr", "localhost:8888")
 
 For more information on IMD v2 as it is implemented in GROMACS, see the imd command line
 arguments described `here <https://manual.gromacs.org/documentation/5.1/onlinehelp/gmx-mdrun.html>`_,
@@ -81,7 +80,7 @@ def __init__(
         self,
         filename,
         convert_units=True,
-        num_atoms=None,
+        n_atoms=None,
         buffer_size=2**26,
         socket_bufsize=None,
         **kwargs,
@@ -98,9 +97,9 @@ def __init__(
 
         # NOTE: Replace me with header packet which contains this information
         # OR get this information from the topology?
-        if not num_atoms:
-            raise ValueError("num_atoms must be specified")
-        self.n_atoms = num_atoms
+        if not n_atoms:
+            raise ValueError("`n_atoms` kwarg must be specified")
+        self.n_atoms = n_atoms
         self.ts = self._Timestep(
             self.n_atoms, positions=True, **self._ts_kwargs
         )

imdreader/tests/test_imdreader.py

@@ -156,7 +156,6 @@ def test_traj_len(run_gmx):
     u = mda.Universe(
         IMDGROUP_GRO,
         f"localhost:{port}",
-        num_atoms=100,
     )
     for ts in u.trajectory:
         pass
@@ -174,7 +173,6 @@ def test_pause(run_gmx, caplog):
     u = mda.Universe(
         IMDGROUP_GRO,
         f"localhost:{port}",
-        num_atoms=100,
         # 1240 bytes per frame
         buffer_size=62000,
     )
@@ -194,7 +192,6 @@ def test_no_connection(caplog):
     u = mda.Universe(
         IMDGROUP_GRO,
         "localhost:8888",
-        num_atoms=100,
         buffer_size=62000,
     )
     for ts in u.trajectory: