#!/usr/bin/env python3 # Copyright © 2014 Mattias Andrée (maandree@member.fsf.org) # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU Affero General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU Affero General Public License for more details. # # You should have received a copy of the GNU Affero General Public License # along with this program. If not, see . # This module implements support for ICC profiles import os from subprocess import Popen, PIPE from curve import * LIBEXECDIR = 'bin' ''' :str Path to executable libraries, '/usr/libexec' is standard ''' def load_icc(pathname): ''' Load ICC profile from a file @param pathname:str The ICC profile file @return :()→void Function to invoke, parameterless, to apply the ICC profile to the colour curves ''' with open(pathname, 'rb') as file: return parse_icc(file.read()) def get_current_icc(display = None): ''' Get all currently applied ICC profiles as profile applying functions @param display:str? The display to use, `None` for the current one @return list<(screen:int, monitor:int, profile:()→void)> List of used profiles ''' return [(screen, monitor, parse_icc(profile)) for screen, monitor, profile in get_current_icc_raw(display)] def get_current_icc_raw(display = None): ''' Get all currently applied ICC profiles as raw profile data @param display:str? The display to use, `None` for the current one @return list<(screen:int, monitor:int, profile:bytes())> List of used profiles ''' # Generate command line arguments to execute command = [LIBEXECDIR + os.sep + 'blueshift_iccprofile'] if display is not None: command.append(display) # Spawn the libexec blueshift_iccprofile process = Popen(command, stdout = PIPE) # Wait for the child process to exit and gather its output to stdout lines = process.communicate()[0].decode('utf-8', 'error').split('\n') # Ensure the tha process has exited while process.returncode is None: process.wait() # Throw exception if the child process failed if process.returncode != 0: raise Exception('blueshift_iccprofile exited with value %i' % process.returncode) rc = [] # Get the screen, output and profile for each monitor with an _ICC_PROFILE(_n) atom set for s, m, p in [line.split(': ') for line in lines if not line == '']: # Convert the program from hexadecminal encoding to raw octet encoding p = bytes([int(p[i : i + 2], 16) for i in range(0, len(p), 2)]) # List the profile rc.append((int(s), int(m), p)) return rc def parse_icc(content): ''' Parse ICC profile from raw data @param content:bytes The ICC profile data @return :()→void Function to invoke, parameterless, to apply the ICC profile to the colour curves ''' # Magic number for dual-byte precision lookup table based profiles MLUT_TAG = 0x6d4c5554 # Magic number for gamma–brightness–contrast based profiles # and for variable precision lookup table profiles VCGT_TAG = 0x76636774 def fcurve(R_curve, G_curve, B_curve): ''' Apply an ICC profile mapping @param R_curve:list Lookup table for the red channel @param G_curve:list Lookup table for the green channel @param B_curve:list Lookup table for the blue channel ''' for curve, icc in curves(R_curve, G_curve, B_curve): for i in range(i_size): # Nearest neighbour y = int(curve[i] * (len(icc) - 1) + 0.5) # Trunction to actual neighbour y = min(max(0, y), len(icc) - 1) # Apply mapping curve[i] = icc[y] # Integers in ICC profiles are encoded with the most significant byte first int_ = lambda bs : sum([v << (i * 8) for i, v in enumerate(reversed(bs))]) def read(n): ''' Read a set of bytes for the encoded ICC profile @param n:int The number of bytes to read @return :list The next `n` bytes of the profile ''' if len(content) < n: raise Exception('Premature end of file: %s' % pathname) rc, content[:] = content[:n], content[n:] return rc # Convert profile encoding format for bytes to integer list content = list(content) # Skip the first 128 bytes read(128) # Get the number of bytes n_tags, ptr = int_(read(4)), 128 + 4 # Create array for the lookup tables to create R_curve, G_curve, B_curve = [], [], [] for i_tag in range(n_tags): # Get profile encoding type, offset to the profile and the encoding size of its data (tag_name, tag_offset, tag_size), ptr = [int_(read(4)) for _ in range(3)], ptr + 3 * 4 # XXX should I not jump to the data now instead of inside the if statements' bodies? if tag_name == MLUT_TAG: ## The profile is encododed as an dual-byte precision lookup table # Jump to the profile data read(tag_offset - ptr) # Get the lookup table for the red channel, for _ in range(256): R_curve.append(int_(read(2)) / 65535) # for the green channel for _ in range(256): G_curve.append(int_(read(2)) / 65535) # and for the blue channel. for _ in range(256): B_curve.append(int_(read(2)) / 65535) return lambda : fcurve(R_curve, G_curve, B_curve) elif tag_name == VCGT_TAG: ## The profile is encoded as with gamma, brightness and contrast values # or as a variable precision lookup table profile # Jump to the profile data read(tag_offset - ptr) # VCGT profiles starts where their magic number tag_name = int_(read(4)) if not tag_name == VCGT_TAG: break # Skip four bytes read(4) # and get the actual encoding type gamma_type = int_(read(4)) if gamma_type == 0: ## The profile is encoded as a variable precision lookup table (n_channels, n_entries, entry_size) = [int_(read(2)) for _ in range(3)] if tag_size == 1584: (n_channels, n_entries, entry_size) = 3, 256, 2 if not n_channels == 3: # Assuming sRGB, can only be an correct assumption if there are exactly three channels break # Calculate the divisor for mapping to [0, 1] divisor = (256 ** entry_size) - 1 # Values are encoded in integer form with `entry_size` bytes int__ = lambda : int_(read(entry_size)) / divisor # Get the lookup table for the red channel, for _ in range(n_entries): R_curve.append(int__()) # for the green channel for _ in range(n_entries): G_curve.append(int__()) # and for the blue channel. for _ in range(n_entries): B_curve.append(int__()) return lambda : fcurve(R_curve, G_curve, B_curve) elif gamma_type == 1: ## The profile is encoded with gamma, brightness and contrast values # Get the gamma, brightness and contrast for the red channel, (r_gamma, r_min, r_max) = [int_(read(4)) / 65536 for _ in range(3)] # green channel (g_gamma, g_min, g_max) = [int_(read(4)) / 65536 for _ in range(3)] # and blue channel. (b_gamma, b_min, b_max) = [int_(read(4)) / 65536 for _ in range(3)] def f(): ''' Apply the gamma, brightness and contrast ''' # Apply gamma gamma(r_gamma, g_gamma, b_gamma) # before brightness and contrast rgb_limits(r_min, r_max, g_min, g_max, b_min, b_max) return f break # XXX should I not jump to (tag_offset + tag_size - ptr) here # and not break the loops when unknown? raise Exception('Unsupported ICC profile file') def make_icc_interpolation(profiles): ''' An interpolation function for ICC profiles @param profiles:list<()→void> Profile applying functions @return :(timepoint:float, alpha:float)→void() A function that applies an interpolation of the profiles, it takes to arguments: the timepoint and the filter alpha. The timepoint is normally a [0, 1] floating point of the dayness level, this means that you only have two ICC profiles, but you have multiple profiles, in such case profile #⌊timepoint⌋ and profile #(⌊timepoint⌋ + 1) (modulus the number of profiles) are interpolated with (timepoint - ⌊timepoint⌋) weight to the second profile. The filter alpha is a [0, 1] floating point of the degree to which the profile should be applied. ''' def f(t, a): # Get floor and ceiling profiles and weight (pro0, pro1), t = [profiles[(int(t) + 0) % len(profiles)] for i in range(2)], t % 1 if (pro0 is pro1) and (a == 1): # If the floor and ceiling are the same profile, # and the alpha is 1, than we can just simple apply # one without any interpolation pro0() return # But otherwise, we will need to save the current curves r_, g_, b_ = r_curve[:], g_curve[:], b_curve[:] # reset the curves start_over() # and apply on of the profiles pro0() # so that we can get the mapping of one of the profiles. r0, g0, b0 = r_curve[:], g_curve[:], b_curve[:] # After which we can get the last encoding value n = len(r0) - 1 rgb = None if pro0 is pro1: # Now, if the floor and ceiling profiles are than same profile, # then we can use just one of then interpolat between it and a clean adjustment. rgb = [[v * a + i * (1 - a) / n for i, v in enumerate(c0)] for c0 in (r0, g0, b0)] else: # Otherwise we need to clear the curves from the floor profile's adjustments start_over() # and apply the ceiling profile pro1() # so that we can that profile's adjustments. # Than we pair the floor and ceilings profile for each channel. r01, g01, b01 = (r0, r_curve[:]), (g0, g_curve[:]), (b0, b_curve[:]) # Now that we have two profiles, when we interpolate between them and a clean # state, we first interpolate the profiles and the interpolate between that # interpolation and a clean adjustment. interpol = lambda i, v0, v1 : (v0 * (1 - t) + v1 * t) * a + i * (1 - a) / n rgb = [[interpol(i, v0, v1) for i, (v0, v1) in enumerate(zip(*c01))] for c01 in (r01, g01, b01)] # Now that we have read the profiles, it is time to restore the curves to the # state they were in, r_curve[:], g_curve[:], b_curve[:] = r_, g_, b_ # and apply the interpolated profile adjustments on top of it. for curve, icc in curves(*rgb): for i in range(i_size): # Nearest neighbour y = int(curve[i] * (len(icc) - 1) + 0.5) # Trunction to actual neighbour y = min(max(0, y), len(icc) - 1) # Apply mapping curve[i] = icc[y] return f