update docs

doc/reference/input.rst

@@ -12,16 +12,21 @@ The input file consists of six parameter blocks named [model], [system], [impuri
 The following table shows which blocks are used by each program.
 
 .. csv-table::
-    :header: "", ``dcore_pre``, ``dcore``, ``dcore_check``, ``dcore_post``
-    :widths: 5, 5, 5, 5, 5
-
-    [model]          ,  Yes, Yes, Yes, Yes
-    [pre]            ,  Yes,    ,    ,
-    [system]         ,     , Yes, Yes, Yes
-    [impurity_solver],     , Yes,    , Yes
-    [control]        ,     , Yes
-    [tool]           ,     ,    , Yes, Yes
-    [mpi]            ,     , Yes,    , Yes
+    :header: "", ``dcore_pre``, ``dcore``, ``dcore_check``, ``dcore_anacont``, ``dcore_spectrum``
+    :widths: 5, 5, 5, 5, 5, 5
+
+    [model]             ,  Yes, Yes, Yes, Yes, Yes
+    [pre]               ,  Yes,    ,    ,    ,
+    [system]            ,     , Yes, Yes, Yes, Yes
+    [impurity_solver]   ,     , Yes,    ,    , Yes
+    [control]           ,     , Yes
+    [post]              ,     ,    , Yes, Yes, Yes
+    [post.check]        ,     ,    , Yes,    , 
+    [post.anacont]      ,     ,    ,    , Yes, 
+    [post.anacont.pade] ,     ,    ,    , Yes, 
+    [post.anacont.spm]  ,     ,    ,    , Yes, 
+    [post.spectrum]     ,     ,    ,    ,    , Yes
+    [mpi]               ,     , Yes,    ,    , Yes
 
 For example, we can see that ``dcore_pre`` reads only [model] and [pre] blocks. Therefore, ``dcore_pre`` needs to be re-executed only when the parameters in [model] and [pre] blocks are changed.
 
@@ -112,14 +117,49 @@ This block includes parameters that control the self-consistency loop of DMFT.
 
 .. include:: control_desc.txt
 
-[tool] block
+[post] block
 ------------
 
-This block includes parameters that are solely used by ``dcore_post``.
+This block includes parameters that are solely used by ``dcore_anacont`` and ``dcore_spectrum``.
 
 .. ``dcore_check`` and ``dcore_post`` read this block.
 
-.. include:: tool_desc.txt
+.. include:: post_desc.txt
+
+[post.anacont] block
+-----------------------
+
+This block includes parameters that are solely used by ``dcore_anacont``.
+
+.. include:: post.anacont_desc.txt
+
+[post.anacont.pade] block
+---------------------------
+
+This block includes parameters that are solely used by ``dcore_anacont``.
+
+.. include:: post.anacont.pade_desc.txt
+
+[post.anacont.spm] block
+---------------------------
+
+This block includes parameters that are solely used by ``dcore_anacont``.
+
+.. include:: post.anacont.spm_desc.txt
+
+[post.spectrum] block
+-----------------------
+
+This block includes parameters that are solely used by ``dcore_spectrum``.
+
+.. include:: post.spectrum_desc.txt
+
+[post.check] block
+-----------------------
+
+This block includes parameters that are solely used by ``dcore_check``.
+
+.. include:: post.check_desc.txt
 
 [mpi] block
 ------------

doc/tools/akw_converter/akw.rst

@@ -7,15 +7,22 @@ An auxiliary tool ``akw_converter`` converts these data into three-dimensional v
 How to use
 ----------
 
-The following steps should be done after all calculations are finished, namely, self-consistent calculation by ``dcore`` and even post calculations by ``dcore_post``.
+The following steps should be done after all calculations are finished, namely, self-consistent calculation by ``dcore`` and even post calculations by ``dcore_spectrum``.
 
 #. **Modifying ini file**
 
-   A parameter ``nk_mesh`` is added to the ini file
+   To change the output directory from the default value (``post``), ``dir_post`` is added to the ini file as
 
    ::
 
-     [tools]
+     [post]
+     dir_post = post_grd
+
+   Then, to activate the grid calculation, the following block is added to the ini file
+
+   ::
+
+     [post.spectrum]
      nk_mesh = 100
 
    This parameter activates a grid calculation.
@@ -24,7 +31,7 @@ The following steps should be done after all calculations are finished, namely,
 
    ::
 
-     [tools]
+     [post.spectrum]
      nk0_mesh = 100
      nk1_mesh = 100
      nk2_mesh = 4
@@ -35,23 +42,24 @@ The following steps should be done after all calculations are finished, namely,
 
    ::
 
+     [post.anacont]
      Nomega = 3
      omega_max = 1.0
      omega_min = -1.0
 
    Then, :math:`\omega=0` and two additional points are selected (``Nomega = 1`` is not allowed).
    In this way, the data points are reduced to :math:`100^2 \times 4 \times 3`.
 
-#. **Executing dcore\_post**
+#. **Executing dcore\_anacont and dcore\_spectrum**
 
-   Now, ``dcore_post`` is (re-)executed by
+   Now, ``dcore_anacont`` and ``dcore_spectrum`` are (re-)executed by
 
    ::
 
-     dcore_post --np 4 --prefix post_grd/ square.ini
+     dcore_anacont square.ini
+     dcore_spectrum --np 4 square.ini
 
    The number of processes in the ``--np`` option  (4 above) should be changed according to your environment.
-   An output directory is specified by the ``--prefix`` option (default is 'post/').
    In this case, the data are saved in 'post_grd' directory.
    We recommend to use this option to have k-path data and k-grid data in separate directories.
 
@@ -67,7 +75,7 @@ The following steps should be done after all calculations are finished, namely,
      akw_converter.py --omega 0.0 post_grd/square_akw_mesh_up.dat post_grd/akw_up.grd
 
    The value of :math:`\omega` is given by the ``--omega`` option. If data do not exist just on this :math:`\omega`, :math:`A(\boldsymbol{k},\omega)` is interpolated using 2 points nearby (depending on the kind of interpolation method).
-   The first argument is the path to the four-dimensional data generated by ``dcore_post``. The second argument is the output file name. The default format is **GRD**, and the output filename should have extension 'grd'.
+   The first argument is the path to the four-dimensional data generated by ``dcore_spectrum``. The second argument is the output file name. The default format is **GRD**, and the output filename should have extension 'grd'.
 
    The full option of ``akw_converter`` can be seen by the command
    ::

doc/tools/akw_converter/dmft_square_akw.ini

@@ -29,7 +29,7 @@ max_step = 20
 sigma_mix = 0.5
 time_reversal = True
 
-[tool]
+[post.spectrum]
 broadening = 0.0
 knode = [(G,0,0,0),(X,0.5,0,0),(M,0.5,0.5,0),(G,0,0,0)]
 nk_line = 100
@@ -41,6 +41,8 @@ nk_line = 100
 nk0_mesh = 100
 nk1_mesh = 100
 nk2_mesh = 4
+
+[post.anacont]
 omega_max = 1.0
 omega_min = -1.0
 Nomega = 3

doc/tutorial/afm/square_afm.ini

@@ -30,11 +30,13 @@ max_step = 30
 sigma_mix = 0.5
 initial_static_self_energy = {0: 'init_se_up.txt', 1: 'init_se_down.txt'}
 
-[tool]
+[post.spectrum]
 broadening = 0.0
 knode = [(G,0.0,0.0,0.0),(X,0.5,0.0,0.0),(M,0.5,0.5,0.0),(G,0.0,0.0,0.0)]
 # knode = [(G,0,0,0),(X,1,0,0),(M,1,1,0),(G,0,0,0)]
 nk_line = 100
+
+[post.anacont]
 omega_max = 6.0
 omega_min = -6.0
 Nomega = 401

doc/tutorial/bethe-t2g/dmft_bethe.ini

@@ -24,7 +24,7 @@ max_step = 40
 sigma_mix = 1.0
 restart = False
 
-[tool]
+[post.anacont]
 omega_max = 15.0
 omega_min =-15.0
 Nomega = 100

doc/tutorial/square/dmft_square_ctseg.ini

@@ -31,10 +31,19 @@ max_step = 20
 sigma_mix = 0.5
 time_reversal = True
 
-[tool]
-broadening = 0.0
-knode = [(G,0.0,0.0,0.0),(X,0.5,0.0,0.0),(M,0.5,0.5,0.0),(G,0.0,0.0,0.0)]
-nk_line = 100
+[post.anacont]
+solver = pade
 omega_max = 6.0
 omega_min = -6.0
 Nomega = 401
+
+[post.anacont.pade]
+n_min = 20
+n_max = 1000
+iomega_max = 1e+20
+eta = 0.1
+
+[post.spectrum]
+knode = [(G,0.0,0.0,0.0),(X,0.5,0.0,0.0),(M,0.5,0.5,0.0),(G,0.0,0.0,0.0)]
+nk_line = 100
+broadening = 0.0

doc/tutorial/square/dmft_square_pomerol.ini

@@ -26,10 +26,19 @@ max_step = 100
 sigma_mix = 0.5
 converge_tol = 1e-5
 
-[tool]
-knode = [(G,0,0,0),(X,0.5,0,0),(M,0.5,0.5,0),(G,0,0,0)]
-nk_line = 100
+[post.anacont]
+solver = pade
 omega_max = 6.0
 omega_min = -6.0
 Nomega = 401
+
+[post.anacont.pade]
+n_min = 20
+n_max = 1000
+iomega_max = 1e+20
+eta = 0.1
+
+[post.spectrum]
+knode = [(G,0,0,0),(X,0.5,0,0),(M,0.5,0.5,0),(G,0,0,0)]
+nk_line = 100
 broadening = 0.4

doc/tutorial/square/square.rst

@@ -97,10 +97,28 @@ For instance, ``check/iter_sigma-ish0.png`` shows how the renormalization factor
 
 The iteration is terminated when the diff (lower figure) reaches ``converge_tol = 1e-5`` at the 13th iteration.
 
-Spectral function : ``dcore_post``
-----------------------------------
-We can calculate real-frequency quantities such as the density of states and the momentum-dependent single-particle excitations using ``dcore_post`` program.
-The analytical continuation from Matsubara frequency to real frequencies is performed using the Pade approximation.
+Analytical continuation of the self-energy : ``dcore_anacont``
+---------------------------------------------------------------------
+
+The self-energy is calculated in the imaginary-time domain in the DMFT loop and saved as ``seedname_sigma_iw.npz``.
+The analytical continuation from Matsubara frequency to real frequencies is required to calculate the spectral function.
+DCore provides a program ``dcore_anacont`` for this purpose.
+Parameters for the analytical continuation are specified in the ``[post.anacont]`` block in the input file.
+``omega_min`` and ``omega_max`` is the minimum and maximum frequency for the output.
+``Nomega`` is the number of frequency points.
+``solver`` is the solver for the analytical continuation; "pade" is the Pade approximation and "spm" is the sparse modeling method.
+Hyperparameters for the solver can be specified in the ``[post.anacont.pade]`` or ``[post.anacont.spm]`` block.
+
+.. code-block:: bash
+
+   $ dcore_anacont dmft_square.ini
+
+The result is stored in ``post/sigma_w.npz``.
+
+Spectral function : ``dcore_spectrum``
+---------------------------------------
+
+After calculating the self-energy on the real-frequency axis, we can also calculate other real-frequency quantities such as the density of states and the momentum-dependent single-particle excitations using ``dcore_spectrum`` program.
 
 .. In the Hubbard-I solver, the self-energy on the real-frequency axis can be directly computed (no analytical continuation is required).
 .. Hence, the impurity problem is solved once more in ``dcore_post``.
@@ -109,18 +127,19 @@ The calculation is done by the following command:
 
 .. code-block:: bash
 
-   $ dcore_post dmft_square.ini --np 1
+   $ dcore_spectrum dmft_square.ini --np 1
 
 After finishing the calculation,
 results are stored in ``post`` directory.
-The data of momentum-resolved spectral functions are output into ``square_akw.dat``.
-We can easily plot the result by using the script file ``square_akw.gp`` for gnuplot:
+The data of momentum-resolved spectral functions are output into ``akw.dat``.
+We can easily plot the result by using the script file ``akw.gp`` for gnuplot:
 
 .. ``square_akw.dat``, ``square_akw.gp`` and ``square_dos.dat`` are generated.
 
 .. code-block:: bash
 
-   $ gnuplot square_akw.gp
+   $ cd post
+   $ gnuplot akw.gp
 
 In the graph shown below, the left and right panels correspond to up-spin and down-spin components, respectively.
 
@@ -131,14 +150,14 @@ In the graph shown below, the left and right panels correspond to up-spin and do
 Here, we have tuned the range of the coloar bar by the command ``set cbrange[0:0.8]`` to get better figure.
 The band width seems reduced than the noninteracting one, 8, but the artificial structure around E=1 and -1 makes it difficult to judge.
 
-The numerical result for the density of states is stored in ``square_dos.dat``.
+The numerical result for the density of states is stored in ``dos.dat``.
 We can plot it using gnuplot as follows:
 
 .. code-block:: gnuplot
 
    set xlabel "Energy"
    set ylabel "DOS"
-   plot "square_dos.dat" w l
+   plot "dos.dat" w l
 
 The result is shown below.
 

doc/tutorial/srvo3/openmx/openmx.rst

@@ -60,14 +60,16 @@ Running self-consistent DMFT calculation: dcore
 
    $ dcore srvo3.ini
 
-Post-processing and data analysis: dcore_post
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Post-processing and data analysis: dcore_anacont and dcore_spectrum
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. code-block :: bash
 
-   $ dcore_post srvo3.ini
-   $ sed -e "s/every 10/every 1/g" srvo3_akw.gp
-   $ gnuplot nis_akw.gp
+   $ dcore_anacont srvo3.ini
+   $ dcore_spectrum srvo3.ini
+   $ cd post
+   $ sed -e "s/every 10/every 1/g" akw.gp
+   $ gnuplot akw.gp
 
 .. image:: akw_srvo3.png
    :width: 500

doc/tutorial/srvo3/openmx/srvo3_openmx.ini

@@ -29,13 +29,18 @@ move_double{bool} = True
 [control]
 max_step = 8
 
-[tool]
+[post.spectrum]
 broadening = 0.1
 nk_line = 50
-nnode = 5
 knode=[(G,0.0,0.0,0.0),(X,0.5,0.0,0.0),(M,0.5,0.5,0.0),(G,0.0,0.0,0.0),(R,0.5,0.5,0.5)]
+
+[post.anacont]
 omega_max =2.0
 omega_min =-2.0
 Nomega = 400
+
+[post.anacont.pade]
+iomega_max = 5.0
+
+[post.check]
 omega_check = 30.0
-omega_pade = 5.0

doc/tutorial/srvo3/srvo3.ini

@@ -28,13 +28,18 @@ max_step = 12
 time_reversal = True
 sigma_mix = 0.8
 
-[tool]
+[post.spectrum]
 broadening = 0.1
 nk_line = 50
-nnode = 5
 knode=[(G,0.0,0.0,0.0),(X,0.5,0.0,0.0),(M,0.5,0.5,0.0),(G,0.0,0.0,0.0),(R,0.5,0.5,0.5)]
+
+[post.anacont]
 omega_max =2.0
 omega_min =-2.0
 Nomega = 400
+
+[post.anacont.pade]
+iomega_max = 2.0
+
+[post.check]
 omega_check = 30.0
-omega_pade = 2.0