bug fixes, documentation in data_management

data_management.py

 """
-Data management for FRTRG data based on pytables
-
-Idea:
-* Data + Metadata are stored in H5F files.
-* A single file can be any logical unit. An arbitrary number of files can be created.
-* Metadata are partially stored in a database for simple access
-* a file "filename.lock" is created before writing to a file.
-
-Copyright 2022 Valentin Bruch
+Kondo FRTRG, data management
+
+This file contains functions and classes to manage data generated using the
+kondo module.
+
+General concepts:
+* All metadata are stored in an SQL database.
+* Floquet matrices are stored in H5F files.
+* Each H5F file can contain multiple data points. Data points can be added to
+  H5F files.
+* Each H5F file contains a table of metadata for the data points stored in
+  this file.
+* Data points are identified in H5F files by a hash generated from their full
+  Floquet matrices at the end of the RG flow.
+* The SQL database stores the directory, filename, and hash where the Floquet
+  matrices are stored.
+
+Implementation:
+* pandas for accessing the SQL database and managing the full table of metadata
+* pytables for H5F files
+* a file "filename.lock" is temporarily created when writing to a H5F file.
+
+Copyright 2022 Valentin Bruch <valentin.bruch@rwth-aachen.de>
 License: MIT
 """
 
@@ -23,11 +37,24 @@ import random
 import settings
 import warnings
 
+# We use hashs as identifiers for data points in H5F files. These hashs are
+# often not valid python names, which causes a warning. We ignore this warning.
 warnings.simplefilter("ignore", tb.NaturalNameWarning)
 
-
-_to_char = lambda x: chr(x + 48 if x < 10 else x + 55 if x < 36 else x + 61)
-random_string = lambda n: ''.join(_to_char(random.randint(0,61)) for i in range(n))
+def random_string(length : int):
+    """
+    Generate random strings of alphanumerical characters with given length.
+    """
+    res = ""
+    for _ in range(length):
+        x = random.randint(0, 61)
+        if x < 10:
+            res += chr(x + 48)
+        elif x < 36:
+            res += chr(x + 55)
+        else:
+            res += chr(x + 61)
+    return res
 
 
 def replace_all(string:str, replacements:dict):
@@ -69,6 +96,7 @@ class KondoExport:
         """
         dictionary of metadata
         """
+        # Determine method
         if self.kondo.unitary_transformation:
             if self.kondo.compact == 2:
                 method = 'J-compact-2'
@@ -78,27 +106,29 @@ class KondoExport:
                 method = 'J'
         else:
             method = 'mu'
+
+        # Collect solver flags
         solver_flags = 0
         try:
             if self.kondo.simplified_initial_conditions:
                 solver_flags |= DataManager.SOLVER_FLAGS["simplified_initial_conditions"]
         except AttributeError:
             pass
-        for (key, value) in kondo.settings.items():
+        for (key, value) in self.kondo.global_settings.items():
             if value:
                 try:
                     solver_flags |= DataManager.SOLVER_FLAGS[key.lower()]
                 except KeyError:
                     pass
 
-        version = kondo.settings["VERSION"]
+        version = self.kondo.global_settings["VERSION"]
         return dict(
                 hash = self.hash,
                 omega = self.kondo.omega,
                 energy = self.kondo.energy,
                 version_major = version[0],
                 version_minor = version[1],
-                lazy_inverse_factor = kondo.settings["LAZY_INVERSE_FACTOR"]
+                lazy_inverse_factor = self.kondo.global_settings["LAZY_INVERSE_FACTOR"],
                 git_commit_count = version[2],
                 git_commit_id = version[3],
                 method = method,
@@ -121,7 +151,12 @@ class KondoExport:
         """
         dictionary of main results: DC current, DC conductance, AC current (absolute value and phase)
         """
-        results = dict(dc_current=np.nan, dc_conductance=np.nan, ac_current_abs=np.nan, ac_current_phase=np.nan)
+        results = dict(
+                dc_current = np.nan,
+                dc_conductance = np.nan,
+                ac_current_abs = np.nan,
+                ac_current_phase = np.nan
+                )
         nmax = self.kondo.nmax
         try:
             results['dc_current'] = self.kondo.gammaL[nmax, nmax].real

gen_data.py

@@ -15,7 +15,7 @@ import settings
 import numpy as np
 
 
-def gen_option_iter(steps, scales, **options):
+def gen_option_iter(steps=None, scales=None, **options):
     """
     Interpret given options to swipe over parameters.
 
@@ -46,9 +46,11 @@ def gen_option_iter(steps, scales, **options):
             options.pop(key)
         steps = [max_length]
     else:
-        if isinstance(scales, str):
+        if scales is None:
+            scales = len(steps) * ["linear"]
+        elif isinstance(scales, str):
             scales = len(steps)*[scales]
-        if len(scales) == 1:
+        elif len(scales) == 1:
             scales *= len(steps)
         for key, value in options.items():
             if type(value) != list or key == "fourier_coef":